Skip to main content
NCBI home page
Wiley - PMC COVID-19 Collection logoLink to Wiley - PMC COVID-19 Collection
. 2022 Oct 7;2(10):e534. doi: 10.1002/cpz1.534

Genome Reporting for Healthy Populations—Pipeline for Genomic Screening from the GENCOV COVID‐19 Study

Erika Frangione 1,2, Monica Chung 1,2, Selina Casalino 1,2, Georgia MacDonald 1,2, Sunakshi Chowdhary 1,2, Chloe Mighton 1,2,3,4, Hanna Faghfoury 5, Yvonne Bombard 3,4, Lisa Strug 6, Trevor Pugh 5,7, Jared Simpson 7, Limin Hao 8, Matthew Lebo 8,9, William J Lane 9, Jennifer Taher 1,3, Jordan Lerner‐Ellis 1,2,3,✉; GENCOV Study Workgroup1
PMCID: PMC9874607  PMID: 36205462

Abstract

Genome sequencing holds the promise for great public health benefits. It is currently being used in the context of rare disease diagnosis and novel gene identification, but also has the potential to identify genetic disease risk factors in healthy individuals. Genome sequencing technologies are currently being used to identify genetic factors that may influence variability in symptom severity and immune response among patients infected by SARS‐CoV‐2. The GENCOV study aims to look at the relationship between genetic, serological, and biochemical factors and variability of SARS‐CoV‐2 symptom severity, and to evaluate the utility of returning genome screening results to study participants. Study participants select which results they wish to receive with a decision aid. Medically actionable information for diagnosis, disease risk estimation, disease prevention, and patient management are provided in a comprehensive genome report. Using a combination of bioinformatics software and custom tools, this article describes a pipeline for the analysis and reporting of genetic results to individuals with COVID‐19, including HLA genotyping, large‐scale continental ancestry estimation, and pharmacogenomic analysis to determine metabolizer status and drug response. In addition, this pipeline includes reporting of medically actionable conditions from comprehensive gene panels for Cardiology, Neurology, Metabolism, Hereditary Cancer, and Hereditary Kidney, and carrier screening for reproductive planning. Incorporated into the genome report are polygenic risk scores for six diseases—coronary artery disease; atrial fibrillation; type‐2 diabetes; and breast, prostate, and colon cancer—as well as blood group genotyping analysis for ABO and Rh blood types and genotyping for other antigens of clinical relevance. The genome report summarizes the findings of these analyses in a way that extensively communicates clinically relevant results to patients and their physicians. © 2022 Wiley Periodicals LLC.

Basic Protocol 1: HLA genotyping and disease association

Basic Protocol 2: Large‐scale continental ancestry estimation

Basic Protocol 3: Dosage recommendations for pharmacogenomic gene variants associated with drug response

Support Protocol: System setup

Keywords: bioinformatics, COVID‐19, genome reporting, genomics, next generation sequencing (NGS)

INTRODUCTION

The implementation of serological and molecular tools to inform COVID‐19 patient management—the GENCOV study (Taher et al., 2021)—incorporates genome sequencing (GS) technologies to identify and characterize genetic factors that may have an impact on symptom severity and immune response in SARS‐CoV‐2‐infected patients. Sequencing of host genomes may provide useful insights for clinical diagnosis, disease prevention, and patient management strategies. In the GENCOV study, medically relevant genetic results and other genetic findings from the analysis of host genome sequencing data are communicated to participants in a comprehensive genome report. The GENCOV study enrolled approximately 1500 patients with a PCR‐positive COVID‐19 nasopharyngeal swab. Blood samples were collected at baseline, 1 month, 6 months, and 1 year after diagnosis. Antibody isotype (IgG, IgA, and IgM), titers, and viral neutralization were analyzed and returned to study participants. DNA used for host genome sequencing was isolated from blood lymphocytes. Viral genome lineages were also sequenced and returned. Pre‐ and post‐test counseling is provided by a certified genetic counsellor to manage clinical referrals by a medical geneticist.

This article describes a bioinformatic pipeline for the purpose of returning genetic results in a genome report. The genome report includes HLA genotyping predictions, an estimation of individual large‐scale continental ancestry, and pharmacogenomic variation for drug response and dosage recommendations. In addition, this report includes existing algorithms to determine patient ABO, Rh, and other blood group genotypes, and an assessment of common disease risk by calculating polygenic risk score (PRS) for six common diseases (atrial fibrillation, coronary artery disease, type 2 diabetes, prostate cancer, colorectal cancer, and breast cancer). The genomes are examined for clinically significant variation related to genetic conditions for the following categories: blood and immunology, endocrine, metabolic/mitochondrial, musculoskeletal, hearing loss, neurology, cardiology, ophthalmology, renal, skin, gastrointestinal disorders, and hereditary cancer. Any clinically significant gene variants detected are assessed using the American College of Medical Genetics (ACMG/AMP) criteria for interpreting sequence variation (Richards et al., 2015). SARS‐CoV‐2 viral lineage was also determined and included on the report (see Fig. 1, and File S1 in Supporting Information).

Figure 1.

Figure 1

Open in a new tab

Components of the GENCOV Comprehensive Genome Report to be shared with participants and family physicians. Elements of the genomic report include basic protocols developed for HLA genotyping and disease association, large‐scale continental ancestry estimation, and pharmacogenomic dosage recommendations. Additional protocols incorporated into the report have been adapted from collaborators, which include PRS score assessment, ABO and Rh blood group genotyping, and viral lineage identification. An additional protocol for determining monogenic disease risk was also developed in‐house and included on the report.

The genomic analyses for this study are being performed in‐house provided through the Canadian Genomics COVID‐19 Network (CanCOGeN) at The Centre for Applied Genomics (TCAG) at the Hospital for Sick Children. Base calling is performed using Illumina HiSeq 2500 to generate standard genome sequencing output in the form of FASTQ read files. After quality control (QC), metrics are applied to validate sequence generation, and the FASTQ reads are aligned to the GRCh37 reference human genome using Burrows‐Wheeler Aligner (BWA; Li & Durbin, 2009). The Genome Analysis Toolkit (GATK) is used to determine copy number variants (CNV) and structural variants (SVs; Mckenna et al., 2010). Detected variants are annotated using a custom pipeline based on Annovar, which is subsequently outputted into a .tsv file format for further analysis. In addition to the standard FASTQ output, Variant Call Format (VCF) files for SNV and indel variants, including SV and CNV calls, are generated through secondary analysis by a third‐party vendor, the Franklin Genoox Platform (https://www.genoox.com). In our current genomic analysis pipelines for the study, FASTQ, VCF, and Binary Alignment Map (BAM) files are used as input.

Basic Protocol 1 explains how to utilize an HLA genotyping software application, HLA‐VBSeq, to generate estimated genotypes for 22 HLA loci, including HLA‐A, ‐B, and ‐C. Using the predicted genotypes from this output, custom python scripts were developed to map the genotypes to an in‐house database containing relevant HLA‐disease associations (Fig. 2). Basic Protocol 2 describes a pipeline for estimating individual large‐scale continental ancestry against a reference population dataset, HapMap3, to generate an individual admixed ancestry report in a graphical representation (Fig. 3). Basic Protocol 3 describes the genotyping of 17 relevant pharmacogenomics genes using the software Stargazer, and maps these identified genotypes to current dosage recommendations from the PharmGKB database (Fig. 4). In order to run all three protocols, users will have the option to git clone from our repository or alternatively access Docker images from DockerHub. The protocols outlined in this pipeline have been validated against the NA12878 genome available through Illumina Platform Pedigree (Eberle et al., 2017). Downloads and generation of this sample data are included as script in each of the Basic Protocol github repositories.

Figure 2.

Figure 2

Open in a new tab

Basic Protocol 1 outlines the pipeline for HLA genotyping and disease association as included on the genome report. This pipeline utilizes the software HLA‐VBSeq for HLA genotype calls in combination with an in‐house‐developed database containing HLA‐disease associations as identified from the literature.

Figure 3.

Figure 3

Open in a new tab

Basic Protocol 2 outlines the pipeline for large‐scale continental ancestry estimation as included on the genome report. This pipeline utilizes the HapMap Phase III reference dataset and ADMIXTURE to determine the admixed ancestral populations across GENCOV participants.

Figure 4.

Figure 4

Open in a new tab

Basic Protocol 3 outlines the pipeline for pharmacogenomics dosage recommendations as included on the genome report. This pipeline utilizes the Stargazer software and formal PharmGKB recommendations to identify pharmacogenomics variants with clinically relevant gene‐drug interactions across GENCOV participants.

Basic Protocol 1. HLA GENOTYPING AND DISEASE ASSOCIATION

Basic Protocol 1 explains how to utilize the HLA‐VBSeq software to generate estimated HLA genotypes for 22 HLA genes, including HLA‐A, HLA‐B, HLA‐C. Using the output from HLA‐VBSeq, python scripts were developed to map predicted genotypes to an in‐house developed database containing relevant HLA‐disease associations.

Necessary Resources

Hardware

This pipeline has been tested on both Linux and Mac OS

Software

Files

  • A sample's BAM, VCF and paired‐end FASTQ files. The BAM and VCF files must be generated from the paired‐end FASTQ read files being aligned to the GRCh37 (or equivalent) reference human genome. In addition, the VCF file must contain SNV and indel variants.

  • 1
    Download the HLA repository from https://github.com/jlelabs/HLA. This repository can be downloaded on Github's web interface or cloned using the git command line.
    • git clone https://github.com/jlelabs/HLA
  • 2
    Download the following files from HLA‐VBSeq at http://nagasakilab.csml.org/hla/ and transfer them into your Configuration folder within the HLA repository.
    • HLAVBSeq.jar
    • parse_result.pl
    • call_hla_digits.py
    • hla_all_v2.fasta
    • Allelelist_v2.txt
  • 3
    Go into the HLA repository using the following command in terminal. The HLA_Repository_Path will be the file path in which the repository was either downloaded or cloned.
    • cd ${HLA_Repository_Path}
  • 4
    To run the HLA pipeline, input the following into terminal:
    • ./HLA.sh ‐i <SAMPLE_ID> ‐v <VCF> ‐b <BAM> ‐‐f1 <FASTQ1> ‐‐f2 <FASTQ2> ‐o <OUTPUT_DIRECTORY>
    There are six mandatory arguments:
    • ‐i: the sample id, or another desired name designated by the user that will be used to name all outputs.
    • ‐v: the path to the sample VCF file.
    • ‐b: the path to the sample BAM file.
    • ‐‐f1: the path to the sample FASTQ read 1 file.
    • ‐‐f2: the path to the sample FASTQ read 2 file.
    • ‐o: the file path of the desired output directory, in which all files will be saved, including final results and script logs.

Basic Protocol 2. LARGE‐SCALE CONTINENTAL ANCESTRY ESTIMATION

Basic Protocol 2 outlines how to estimate individual ancestry against a reference HapMap3 set using PLINK (Purcell et al., 2007) and ADMIXTURE software (Alexander, Novembre, & Lange, 2009). PLINK was used for the filtering and merging, while ADMIXTURE was used to output subject ancestry. In‐house R scripts were developed to produce a visual representation of a given individual's results. Expected outputs for this pipeline can be found in the Guidelines for Understanding Results section.

Necessary Resources

Hardware

This pipeline has been tested on both Linux and Mac OS

Software

Files

  • A sample VCF file containing SNV and indel variants. The VCF file must be generated from the paired end FASTQ read files being aligned to the GRCh37 (or equivalent) reference human genome.

  • 1
    Download the Ancestry repository from https://github.com/jlelabs/Ancestry. This repository can be downloaded on Github's web interface or cloned using the git command line.
    • git clone https://github.com/jlelabs/Ancestry
  • 2
    Go into the Ancestry repository using the following command in terminal. The Ancestry_Repository_Path will be the file path in which the repository was either downloaded or cloned.
    • cd ${Ancestry_Repository_Path}
  • 3
    Run the following command to download the reference HapMap3 set. This will download and configure the needed reference files, and will only have to be run once.
    • ./ancestry_hapmap3.sh
  • 4
    To run the Ancestry pipeline, input the following into terminal.
    • ./ancestry.sh ‐i <SAMPLE_ID> ‐v <VCF> ‐o <OUTPUT_DIRECTORY>
    There are three mandatory arguments:
    • ‐i: the sample id, or another desired name designated by the user that will be used to name all outputs.
    • ‐v: the path to the sample VCF file.
    • ‐o: the file path of the desired output directory, in which all files will be saved, including final results, and script logs.

Basic Protocol 3. DOSAGE RECOMMENDATIONS FOR PHARMACOGENOMIC GENE VARIANTS ASSOCIATED WITH DRUG RESPONSE

Basic Protocol 3 was developed to assign individual metabolizer phenotype status and link appropriate recommendations using bcftools, GATK, and Stargazer. The Stargazer software was used to identify the genotype of 17 pharmacogenes of interest, while an in‐house python script was developed to generate recommendations. The recommendations and metabolizer phenotypes were adapted from the diplotype‐phenotype tables and the gene‐medication dosing guidelines from PharmGKB and the Clinical Pharmacogenetics Implementation Consortium (CPIC). Expected outputs for this pipeline can be found in the Guidelines for Understanding Results section.

Necessary Resources

Hardware

This pipeline has been tested on both Linux and Mac OS

Software

Files

  • A sample's BAM and VCF files. The BAM and VCF files must be generated from the paired end FASTQ read files being aligned to the GRCh37 (or equivalent) reference human genome. In addition, the VCF file must contain SNV and indel variants.

  • 1
    Download the Pharmacogenomics repository from https://github.com/jlelabs/Pharmacogenomics. This repository can be downloaded on Github's web interface or cloned using git command line.
    • git clone https://github.com/jlelabs/Pharmacogenomics
  • 2

    Download the Stargazer software at https://stargazer.gs.washington.edu/stargazerweb/res/form.html and move it into the Configuration folder.

  • 3
    Go into the Pharmacogenomics repository using the following command in terminal. The Pharmacogenomics_Repository_Path will be the file path in which the repository was either downloaded or cloned.
    • cd ${Pharmacogenomics_Repository_Path}
  • 4
    To run the Pharmacogenomics pipeline, input the following into terminal.
    • ./Pharmacogenomics.sh ‐i <SAMPLE_ID> ‐v <VCF> ‐b <BAM> ‐o <OUTPUT_DIRECTORY>
    There are four mandatory arguments:
    • ‐i: the sample id, or another desired name designated by the user that will be used to name all outputs.
    • ‐v: the path to the sample VCF file.
    • ‐b: the path to the sample BAM file.
    • ‐o: the file path of the desired output directory, in which all files will be saved, including final results and script logs.

SYSTEM SETUP

Necessary Resources

Hardware

  • This pipeline has been tested on both Linux and Mac OS. Minimum RAM requirements are 16 GB of RAM and 4 CPUs.

Software

This pipeline requires Docker installation

Files

  • A sample's BAM, VCF, and paired‐end FASTQ files. The BAM and VCF files must be generated from the paired end FASTQ read files being aligned to the GRCh37 (or equivalent) reference human genome. In addition, the VCF file must contain SNV and indel variants.

Using Docker and DockerHub

  • 1

    Download the desired repository from DockerHub.

    Basic Protocol 1:
    • docker pull monicahzchung/hla
    Basic Protocol 2:
    • docker pull monicahzchung/ancestry
    Basic Protocol 3:
    • docker pull monicahzchung/pharma​
  • 2

    To run the images, execute the following commands:

    Basic Protocol 1:
    • docker run ‐v INPUT_DIRECTORY:/INPUT/ \
      • ‐v OUTPUT_DIRECTORY:/OUTPUT \
      • hla \
      • sample_id \
      • /INPUT/VCF \
      • /INPUT/BAM \
      • /INPUT/FASTQ1 \
      • /INPUT/FASTQ2 \
      • /OUTPUT
    Basic Protocol 2:
    • docker run ‐v INPUT_DIRECTORY:/INPUT/ \
      • ‐v OUTPUT_DIRECTORY:/OUTPUT \
      • ancestry \
      • sample_id \
      • /INPUT/VCF \
      • /OUTPUT
    Basic Protocol 3:
    • docker run ‐v INPUT_DIRECTORY:/INPUT/ \
      • ‐v OUTPUT_DIRECTORY:/OUTPUT \
      • pharma \
      • sample_id \
      • /INPUT/VCF \
      • /INPUT/BAM \
      • /OUTPUT

Manually installing dependencies

NOTE: It is recommended to use the most up‐to‐date software installations where possible for the basic protocols outlined above.

GUIDELINES FOR UNDERSTANDING RESULTS

Outputs

Each of the genomic analysis pipelines requires a mandatory output directory in which the user must specify with the ‐o flag. In the specified file path, the following folders and files will be generated. The expected results and directory structure for each of the Basic Protocols are explained in detail below and were validated using the NA12878 genome. The table of possible HLA‐disease associations can be found in File S2 in Supporting Information.

Basic Protocol 1: HLA Genotyping and Disease Association

Folder structure

sample_output_directory/

  • |‐ sample_HLA_REPORT/

    • |‐ sample_allele_table.csv

    • |‐ sample_patient_allele_table.csv

    • |‐ sample_patient_associated_disease.csv

    • |‐ sample_report.d8.txt

    • |‐ HLA.log

    • |‐ sample_HLA_SUPPLEMENTARY/

      • |‐ sample_Allele_Depth.txt

      • |‐ sample_results.txt

      • |‐ sample_Top_2_Results.txt

Sample_HLA_REPORT/ folder:

The HLA_REPORT folder contains information to be shared in the genome reports, as well as the log file for the protocol. The sample_report.d8.txt file is expected output from the HLA‐VBSeq HLA calling software, which estimates most likely HLA alleles to 8‐digit resolution. For the reporting of HLA‐disease risk association, either the sample_patient_allele_table.csv or sample_allele_table.csv can be used for report generation. The sample_allele_table.csv aggregates the alleles by gene, whereas sample_patient_allele_table.csv does not. Sample results from each of the expected outputs are shown below.

The sample_patient_allele_table.csv file reports on the HLA loci of interest, the top two most likely HLA alleles adapted from the sample_report.d8.txt file, a risk column that indicates whether the specific allele(s) are associated with known cases of disease risk as identified in an in‐house developed HLA database, and HLA ID, which is an internal ID that links potential increased risk allele(s) cases to an identified case in the disease risk database. One or more HLA IDs could potentially be assigned to a given HLA locus depending on the number of disease associations identified. If multiple disease associations for a given allele are detected, the internal HLA IDs will be listed together in the HLA ID column. The sample_patient_allele_table.csv file is primarily used to verify the detection of any internal HLA IDs linked to one or more specific alleles.

Report files

Table 1 shows an example of the sample_patient_allele_table.csv file in which a single locus, HLA‐B, was detected as being associated with an increased risk to a specific disease. In this case, the internal ID HLAC20 links the particular HLA‐B*27:04 allele to a potential disease risk of Ankylosing Spondylitis (AS).

Table 1.

Case 1 of sample_patient_allele_table.csv Output

Gene Allele Risk HLA ID
HLA‐A A*02:01:01:01 None None
HLA‐A A*33:03:01:01 None None
HLA‐B B*27:04:01 Risk HLAC20
HLA‐B B*58:01:01:03 None None
HLA‐C C*03:02:02:x1 None None
HLA‐C C*04:01:01:14 None None
HLA‐DPB1 DPB1*02:01:02:29 None None
HLA‐DPB1 DPB1*04:01:34 None None
HLA‐DQA1 DQA1*01:02:01:04 None None
HLA‐DQA1 DQA1*04:01:01:02 None None
HLA‐DQB1 DQB1*04:02:01:04 None None
HLA‐DQB1 DQB1*06:09:01:01 None None
HLA‐DRB1 DRB1*07:01:01:02 None None
HLA‐DRB1 DRB1*08:02:01:01 None None
Open in a new tab

Table 2 shows an example of the sample_patient_allele_table.csv file, in which multiple loci in combination, in this case HLA‐DQA1 and HLA‐DQB1, are detected as having alleles associated with an increased risk for a specific disease. In this case, the internal IDs HLAC6, HLAC7, and HLAC10 link the particular DQA1*05:01 and DQB1*02:01 alleles in combination to a potential disease risk for Celiac Disease (CD).

Table 2.

Case 2 of sample_patient_allele_table.csv Output

Gene Allele Risk HLA ID
HLA‐A A*02:01:01:01 None None
HLA‐A A*33:03:01:01 None None
HLA‐B B*40:01:02:04 None None
HLA‐B B*58:01:01:03 None None
HLA‐C C*03:02:02:x1 None None
HLA‐C C*04:01:01:14 None None
HLA‐DPB1 DPB1*02:01:02:29 None None
HLA‐DPB1 DPB1*04:01:34 None None
HLA‐DQA1 DQA1*05:01:01:01 Risk HLAC7 HLAC10
HLA‐DQA1 DQA1*05:01:01:02 Risk HLAC10
HLA‐DQB1 DQB1*02:01:01 Risk HLAC6 HLAC10
HLA‐DQB1 DQB1*02:01:02 Risk HLAC10
HLA‐DRB1 DRB1*07:01:01:02 None None
HLA‐DRB1 DRB1*08:02:01:01 None None
Open in a new tab

The sample_allele_table.csv file reports on the HLA loci of interest, the top two most likely HLA alleles adapted from the sample_report.d8.txt file, and a risk column that indicates whether the specific allele(s) are associated with known cases of disease risk from the internal HLA database. Note that the HLA alleles in this table only report to 4‐digit resolution, for purposes of simplifying the reporting of allele status to participants.

Table 3 shows an example of the sample_allele_table.csv file in which a single locus, HLA‐B, was detected as being associated with an increased risk to a specific disease (Ankylosing Spondylitis). Disease information on the report will be generated in a separate table.

Table 3.

Case 1 of sample_allele_table.csv Output

Gene Allele 1 Allele 2 Risk
HLA‐A A*02:01 A*33:03 Average risk
HLA‐B B*27:04 B*58:01 Increased risk
HLA‐C C*03:02 C*04:01 Average risk
HLA‐DQA1 DQA1*01:02 DQA1*04:01 Average risk
HLA‐DQB1 DQB1*04:02 DQB1*06:09 Average risk
HLA‐DPB1 DPB1*02:01 DPB1*04:01 Average risk
HLA‐DRB1 DRB1*07:01 DRB1*08:02 Average risk
Open in a new tab

Table 4 shows an example of the sample_allele_table.csv file in which multiple loci in combination, in this case HLA‐DQA1 and HLA‐DQB1, are detected as having alleles associated with an increased risk to a specific disease (Celiac Disease).

Table 4.

Case 2 of sample_allele_table.csv Output

Gene Allele 1 Allele 2 Risk
HLA‐A A*02:01 A*33:03 Average risk
HLA‐B B*40:01 B*58:01 Average risk
HLA‐C C*03:02 C*04:01 Average risk
HLA‐DQA1 DQA1*05:01 DQA1*05:01 Increased risk
HLA‐DQB1 DQB1*02:01 DQB1*02:01 Increased risk
HLA‐DPB1 DPB1*02:01 DPB1*04:01 Average risk
HLA‐DRB1 DRB1*07:01 DRB1*08:02 Average risk
Open in a new tab

The sample_patient_associated_disease.csv file reports on seven columns indicating the internal HLA ID assigned to one or more of the Increased Risk cases within the HLA‐disease risk database, Associated Disease, HLA Locus, Allele, Haplotype or Genotype Assignment, Risk, “OR[95%CI] or RR, p‐value, references”, and Interpretation. The sixth column indicates the HLA‐disease association's odds ratio at a 95% confidence interval, relative risk, and/or p‐value as identified from the literature, including the literature reference that identified a disease association to the specific allele(s) detected. The interpretation column provides in‐depth information regarding the potential disease risk, including a brief summary of the disease, the general population risk in Canada, and any known information about the identified risk alleles. The sample_patient_associated_disease.csv will be blank if all seven HLA loci from the previous two tables report an Average Risk status and do not have an internal HLA ID that matches to a disease association case in the HLA database. Otherwise, if an “Increased Risk” is detected, the associated disease risk information from the HLA database will be populated in the sample_patient_associated_disease.csv table for all specific allele(s) identified. Multiple disease associations may be reported for a single HLA locus depending on genotype assignment.

Table 5 shows an example of the sample_patient_associated_disease.csv file in which a single allele at locus HLA‐B was detected as being associated with an increased risk of a specific disease (Ankylosing Spondylitis). Any known Odds Ratio, Relative Risk, and associated literature references will be presented in the table, including disease information and any known information about the specific risk allele identified.

Table 5.

Case 1 of sample_patient_associated_disease.csv Output

HLA ID Associated disease Locus Allele/haplotype/genotype assignment Risk OR[95%CI] or RR, p‐value, references Interpretation
HLAC20 Ankylosing Spondylitis (AS) B B*27:04 × 1 Increased risk OR = 1.76 (1.20‐2.58), p = .004, study population = meta‐analysis including Asian, African, American, and Europeans (Yang_2013_24261772); RR = 1.14 (1.01‐1.29), p = .041, study population = meta‐analysis (Lin_2017_28526894) AS is a form of inflammatory arthritis that affects the spine and its surrounding joints and ligaments, causing chronic pain. AS usually manifests between 15 and 30 years of age. As many as 1 in 100 Canadians are estimated to live with AS (approximately 1%). The exact cause of AS is currently unknown but is thought to be caused by an interaction of genetic, immunological, and environmental factors. (Zhu_2019_ 31666997) (Arthritis Society, 2019 https://arthritis.ca/about‐arthritis/arthritis‐types‐(a‐z)/types/ankylosing‐spondylitis). As many as 90% of people who develop AS carry an HLA‐B*27 allele; however having the HLA‐B*27 allele is not diagnostic for AS. A positive HLA‐B*27 genetic test result can assist healthcare providers in confirming a diagnosis of AS in the presence of symptoms such as joint pain, stiffness and/or swelling of the spine, neck or chest, as well as inflammation of the urethra and eye (Arthritis Society, 2019; https://arthritis.ca/about‐arthritis/arthritis‐types‐(a‐z)/types/ankylosing‐spondylitis)
Open in a new tab

Table 6 shows an example of the sample_patient_associated_disease.csv file in which alleles in combination at multiple loci, HLA‐DQA1 and HLA‐DQB1, were detected as being associated with an increased risk of a specific disease (Celiac Disease). Any known Odds Ratio, Relative Risk, and associated literature references will be presented in the table, including disease information and any known information about the specific risk allele identified.

Table 6.

Case 2 of sample_patient_associated_disease.csv Output

HLA ID Associated disease Locus Allele/haplotype/genotype assignment Risk OR[95%CI] or RR, p‐value, references Interpretation
HLAC6 Celiac Disease (CD) DQB1 DQB1*02:01×1 Increased risk OR = 20 (8.75‐40.73), p < .001, study population = Slovenian (Schweiger_2016_27138053) CD is an autoimmune disorder wherein the surface of the small intestine is damaged by gluten (a group of proteins present in wheat, rye, barley, spelt, kamut, and other various cereals), causing gastrointestinal issues like diarrhea, constipation, and bloating. The ingestion of gluten‐containing foods is a known trigger of CD symptoms. CD is estimated to affect about 1 in 114 Canadians, and 0.5%‐1% of the general population. CD is thought to be caused by a combination of genetic, immunological, and environmental factors (Caio_2019_ 31331324) (Health Canada, 2018; https://www.canada.ca/en/health‐canada/services/food‐nutrition/reports‐publications/food‐safety/celiac‐disease‐gluten‐connection‐1.html; Jamnik et al., 2017; https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5640059/). Genetic testing for the HLA‐DQA1 and HLA‐DQB1 alleles may be used to predict an individual's risk of developing CD. Individuals with HLA‐DQA1 or HLA‐DQB1 are estimated to have a 36%‐53% risk of developing CD. These genes are also quite common in the general population and, therefore, are necessary but not sufficient for the development of CD (Martina_2018_ 30561391).
HLAC7 Celiac Disease (CD) DQA1 DQA1*05:01×1 Increased risk OR = 9.43 (4.52‐19.81), p < .001, study population = Slovenian (Schweiger_2016_27138053) CD is an autoimmune disorder wherein the surface of the small intestine is damaged by gluten (a group of proteins present in wheat, rye, barley, spelt, kamut, and other various cereals), causing gastrointestinal issues like diarrhea, constipation, and bloating. Thengestionn of gluten‐containing foods is a known trigger of CD symptoms. CD is estimated to affect about 1 in 114 Canadians, and 0.5%‐1% of the general population. CD is thought to be caused by a combination of genetic, immunological, and environmental factors (Caio_2019_ 31331324) (Health Canada, 2018 https://www.canada.ca/en/health‐canada/services/food‐nutrition/reports‐publications/food‐safety/celiac‐disease‐gluten‐connection‐1.html; Jamnik et al. 2017; https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5640059/). Genetic testing for the HLA‐DQA1 and HLA‐DQB1 alleles may be used to predict an individual's risk of developing CD. Individuals with HLA‐DQA1 or HLA‐DQB1 are estimated to have a 36%‐53% risk of developing CD. These genes are also quite common in the general population and, therefore, are necessary but not sufficient for the development of CD (Martina_2018_ 30561391).
HLAC10 Celiac Disease (CD) DQA1‐DQB1 DQA1*05:01(or *05:02)×1‐DQB1*02:02(or *02:01)×1/DQA1*02×1‐DQB1*02:02×1 Increased risk OR = 19.68 (2.44‐158.91), p < .001, study population = Slovenian (Schweiger_2016_27138053); OR = 15.95 (2.08‐121.8), p = .0004, study population = Brazilian (Mixed) (Almeida_2016_28042478); Risk = 1 : 7, study population = Brazilian (Mixed) (Almeida_2016_28042478); OR = 8.32 (3.02‐22.93), p < .01, study population = Syrian (Murad_2018_29793442); OR = 9.19 (2.34‐20.14), p < 10−7, study population = Spanish (Martínez‐Ojinaga_2018_29699404) CD is an autoimmune disorder wherein the surface of the small intestine is damaged by gluten (a group of proteins present in wheat, rye, barley, spelt, kamut, and other various cereals), causing gastrointestinal issues like diarrhea, constipation, and bloating. The ingestion of gluten‐containing foods is a known trigger of CD symptoms. CD is estimated to affect about 1 in 114 Canadians, and 0.5%‐1% of the general population. CD is thought to be caused by a combination of genetic, immunological, and environmental factors (Caio et al., 2019 31331324) (Health Canada, 2018; https://www.canada.ca/en/health‐canada/services/food‐nutrition/reports‐publications/food‐safety/celiac‐disease‐gluten‐connection‐1.html; Jamnik et al. 2017; https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5640059/). Genetic testing for the HLA‐DQA1 and HLA‐DQB1 alleles may be used to predict an individual's risk of developing CD. Individuals with HLA‐DQA1 or HLA‐DQB1 are estimated to have a 36%‐53% risk of developing CD. These genes are also quite common in the general population and, therefore, are necessary but not sufficient for the development of CD (Martina_2018_ 30561391).
Open in a new tab

Sample_HLA_Supplementary/ folder

The HLA_SUPPLEMENTARY folder contains supplementary information that is not shared in genome reports, including all outputs generated as part of the HLA‐VBSeq algorithm. These files include a sample_results.txt file, which is a tab‐delimited text file containing the following columns: HLA allele ID, genomic locus length, number of reads that the algorithm assigned to the HLA allele, normalized number of reads, and relative abundance. The sample_Allele_Depth.txt file is a tab‐delimited text file with the following columns: HLA allele name, and an average depth of coverage. The sample_Top_2_Results.txt file is a tab‐delimited text file that reports the top two average depth of coverage calls for 7 HLA loci of interest: HLA‐A, ‐B, ‐C, ‐DQA1, ‐DQB1, ‐DRB1, ‐DPB1). These results are adapted from the sample_Allele_Depth.txt file.

Basic Protocol 2: Large‐Scale Continental Ancestry Estimation

Folder structure

sample_output_directory/

  • |‐ sample_ANCESTRY_REPORT/

    • |‐ sample_HapMap3.png

    • |‐ sample_HapMap3.txt

  • |‐ sample_ANCESTRY_SUPPLEMENTARY/

    • |‐ sample_ADMIXTURE_ref_HAPMAP3.11.P

    • |‐ sample_ADMIXTURE_ref_HAPMAP3.11.Q

    • |‐ sample.pruned.fam

    • |‐ log/

      • |‐ all log files from PLINK

Sample_ANCESTRY_REPORT/ folder

The ANCESTRY_REPORT folder contains information to be included in the genome reports, similar to the HLA_REPORT folder. The sample_HapMap3.png file is a graphical visualization output from the in‐house‐developed R script, taking inputs from ADMIXTURE and PLINK. The ADMIXTURE software estimates large continental ancestry for a given sample, and the .fam file generated from PLINK is used to map each of the mapped references to their designated populations from the HapMap3 reference database (The International HapMap3 Consortium et al., 2010) HapMap3. The results are visualized using the R packages tidyverse, dplyr, RColorBrewer, and reshape2. For reporting, sample_HapMap3.png and sample_HapMap3.txt are used as the figure and caption, respectively.

Sample_ANCESTRY_SUPPLEMENTARY/ folder

The ANCESTRY_SUPPLEMENTARY folder contains supplementary information not shared in genome reports, similar to the sample_HLA_SUPPLEMENTARY folder. The files included in the folder include the .P and .Q file outputs from the ADMIXTURE software and the .fam PLINK binary file generated from the final filtered merged dataset. The .P and .Q files are the allele frequencies of the inferred ancestral populations and the ancestry fractions respectively, while the .fam file is a tab‐delimited text file containing the columns family ID, within‐family ID, within‐family ID of father, within‐family ID of mother, sex code, and phenotype value, if available. In addition, the log folder is also located in this supplementary folder, containing all the log files associated with each of the PLINK operations, as well as the pipeline log file. The sample_HapMap3.png shows a graphical representation of the participant's admixed ancestry as a pie chart (Fig. 5).

Figure 5.

Figure 5

Open in a new tab

Combined pie chart and associated caption summarizing a given sample's estimated ancestry based on HapMap3 reference data. In the pie chart, estimated ancestries are represented by a percentage, with a threshold of equal to or greater than 3% admixed ancestry. Estimated ancestries of less than 3% are not depicted on the pie chart, but indicated instead in the chart legend.

Basic Protocol 3: Dosage Recommendations for Pharmacogenomic Gene Variants Associated With Drug Response

Folder structure

sample_output_directory/

  • |‐ sample_PHARMACOGENOMICS_REPORT/

    • |‐ sample_appendix.xlsx

    • |‐ sample_report.xlsx

    • |‐ Pharmacogenomics.log

  • |‐ sample_PHARMACOGENOMICS_SUPPLEMENTARY/

    • |‐ sample_final_stargazer_output.txt

    • |‐ sample_rsID_Warfarin.txt

    • |‐ sample_Target_GDF_RYR1.gdf (and associated files)

Sample_PHARMACOGENOMICS_REPORT/ folder

The PHARMACOGENOMICS_REPORT folder contains curated pharmacogenomics information to be included in the genome reports, similarly to the HLA_REPORT and ANCESTRY_REPORT folders. This folder contains the appendix.xlsx and report.xlsx files, as well as the Pharmacogenomics.log pipeline log file. The report.xlsx file contains the three columns Gene/Genotype/rsID, Medication(s), and Metabolizer Phenotype. For the Gene/Genotype/rsID column, a gene, its predicted genotype, and associated rsIDs are listed. In the Medication(s) column, the list of medications possibly impacted by the variants present in the gene are provided, while the Metabolizer Phenotype column describes the potential metabolic activity of the gene variants associated with the listed medications for the given individual. The list of all possible pharmacogenes and associated recommendations can be found in File S3 in Supporting Information.

Report files

General metabolizer phenotype classification

Table 7 shows an example report of an individual who has intermediate metabolism for the genes CYP2C19, CYP2C9, an unknown metabolizer status for CYP2D6 and UGT1A1, and poor metabolism for CYP3A5. The specific genotype of CYP2C9 indicates a poor metabolizer status for the medication warfarin, including a specific variant rsID located in the CYP2C chromosome 10 region that is also known to be associated with poor warfarin metabolism (Whirl‐Carrillo et al., 2021).

Table 7.

Sample_report.xlsx

Gene genotype rsID Medication(s) Metabolizer phenotype

CYP2C19

*1/*2

rs12769205

Amitriptyline

Citalopram

Escitalopram

Clomipramine

Clopidogrel

Dexlansoprazole

Doxepin

Imipramine

Lansoprazole

Omeprazole

Pantoprazole

Sertraline

Trimipramine

Voriconazole

 Intermediate metabolizer

CYP2C9

*1/*2

rs1799853

Celecoxib

Flurbiprofen

Ibuprofen

Lornoxicam

Fosphenytoin

Phenytoin

Meloxicam

Piroxicam

Tenoxicam

 Intermediate metabolizer

CYP2D6

*3/*68+*4

rs35742686

rs3892097

Amitriptyline

Atomoxetine

Clomipramine

Codeine

Desipramine

Doxepin

Fluvoxamine

Hydrocodone

Imipramine

Nortriptyline

Ondansetron

Paroxetine

Tamoxifen

Tramadol

Trimipramine

Tropisetron

 Unknown

CYP3A5

*3/*3

rs776746

rs776746

 Tacrolimus Poor metabolizer

UGT1A1

*1/*60

rs4124874

 Atazanavir Unknown

CYP2C9

*1/*2

rs1799853

 Warfarin Poor function

rs12777823

A/G

 Warfarin Poor function
Open in a new tab

The appendix.xlsx file contains a more detailed version of report.xlsx, containing the six following columns: Gene/Genotype/rsID/PharmGKB ID, Medication(s), Metabolizer Phenotype, Implications, Recommendation, Strength of Recommendation. The first three columns in appendix.xlsx are identical to the columns represented in report.xlsx, except that the Gene/Genotype/rsID/PharmGKB ID column also includes the PharmGKB ID for the given clinical and dosage recommendations. Implications for the individual's metabolizer phenotype for the listed medications are detailed, and a dosage recommendation is also assigned. The strength of recommendation is also included as indicated from the PharmGKB/CPIC Database (Whirl‐Carrillo et al., 2021).

Table 8 shows an example of the Appendix file for an individual who has an uncertain susceptibility for malignant hyperthermia based on their CACNA1S genotype, and has a CFTR genotype that is non‐responsive to the medication Ivacaftor in cystic fibrosis patients. The implications of these gene‐drug interactions, the dosage recommendations, and the strength of the recommendations are indicated according to current PharmGKB guidelines (Whirl‐Carrillo et al., 2021). Additional rows for the Appendix are omitted from this table due to the length. The full example output is available as part of File S4 in Supporting Information.

Table 8.

Sample_appendix.xlsx

GeneGenotypersIDPharmGKB ID Medication(s) Metabolizer phenotype Implications Recommendation Strength of recommendation

CACNA1S

Reference/Reference

PA166180457

Desflurane

Enflurane

Halothane

Isoflurane

Methoxyflurane

Sevoflurane

Succinylcholine

 Uncertain Susceptibility These results do not eliminate the chance that this patient is susceptible to Malignant Hyperthermia. The genetic cause of about half of all MH survivors, with MH susceptibility confirmed by contracture test, remains unknown Clinical findings, family history, further genetic testing and other laboratory data should guide use of halogenated volatile anesthetics or depolarizing muscle relaxants Strong

CFTR

*1/*1

PA166114461

 Ivacaftor Ivacaftor non‐responsive in CF patients Ivacaftor treatment is recommended only in cystic fibrosis (CF) patients that are either homozygous or heterozygous for certain CFTR variants. Ivacaftor not recommended Moderate
Open in a new tab

The metabolizer phenotype structure of a typical gene‐drug association is shown in Table 9 for CYP2B6 and Efavirenz. Generally, a given gene variant will report either a Normal Metabolizer, Intermediate Metabolizer, Poor Metabolizer, Rapid Metabolizer, or Ultrarapid Metabolizer phenotype based on genotype. The assignment of metabolizer phenotypes is adapted from the PharmGKB diplotype‐phenotype tables. Each of these possible phenotypes has specific implications, dosage recommendations, and a strength of recommendation for a given gene. This may include a recommendation for increasing or decreasing dosages, or an alternative treatment plan.

Table 9.

Adapted Dosage Recommendation Table

Gene Medication(s) Metabolizer phenotype Implications Recommendation Strength of recommendation PharmGKB ID
CYP2B6 Efavirenz Ultrarapid metabolizer Slightly lower dose‐adjusted trough concentrations of efavirenz compared with normal metabolizers Initiate efavirenz with standard dosing (600 mg/day) Strong PA166182603
CYP2B6 Efavirenz Rapid metabolizer Slightly lower dose‐adjusted trough concentrations of efavirenz compared with normal metabolizers Initiate efavirenz with standard dosing (600 mg/day) Strong PA166182603
CYP2B6 Efavirenz Normal metabolizer Normal efavirenz metabolism Initiate efavirenz with standard dosing (600 mg/day) Strong PA166182603
CYP2B6 Efavirenz Intermediate metabolizer Higher dose‐adjusted trough concentrations of efavirenz compared with normal metabolizers; increased risk of CNS adverse events Consider initiating efavirenz with decreased dose of 400 mg/day Moderate PA166182603
CYP2B6 Efavirenz Poor metabolizer Higher dose‐adjusted trough concentrations of efavirenz compared with normal metabolizers; significantly increased risk of CNS adverse events and treatment discontinuation Consider initiating efavirenz with decreased dose of 400 or 200 mg/day Moderate PA166182603
Open in a new tab

Metabolizer Phenotype Classification—Unique Cases

Some genes are associated with specific metabolizer statuses that differ significantly from the typical phenotype classifications, and are assigned based on specific dosage guidelines. In these unique cases, a metabolizer phenotype may be associated with a certain disease or condition susceptibility, such as in the case of RYR1 and CACNA1S (Table 10). The assignment of the metabolizer phenotypes for RYR1 and CACNA1S are based on the malignant hyperthermia causative variant list (Whirl‐Carrillo et al., 2021). If a given individual is heterozygous for a malignant hyperthermia variant, they will be assigned the metabolizer phenotype of Malignant Hyperthermia Susceptible, or will otherwise be reported as Uncertain Susceptibility.

Table 10.

RYR1 and CACNA1S Metabolizer Phenotype Classifications

Gene Medication(s) Genotype Variant Metabolizer phenotype Implications Recommendation Strength of recommendation PharmGKB ID

RYR1

CACNA1S

Desflurane

Enflurane

Halothane

Isoflurane

Methoxyflurane

Sevoflurane

Succinylcholine

 An individual heterozygous for an RYR1 or CACNA1S malignant hyperthermia causative variant as designated by the EMHG Causative variant Malignant hyperthermia susceptible Individuals are at increased risk of developing malignant hyperthermia if administered halogenated volatile anesthetics or the depolarizing muscle relaxant succinylcholine Halogenated volatile anesthetics or depolarizing muscle relaxants succinylcholine are relatively contraindicated in persons with MHS. They should not be used, except in extraordinary circumstances where the benefits outweigh the risks. In general, alternative anesthetics are widely available and effective in patients with MHS Strong PA166180457

RYR1

CACNA1S

Desflurane

Enflurane

Halothane

Isoflurane

Methoxyflurane

Sevoflurane

Succinylcholine

 An individual negative for a RYR1 or CACNA1S malignant hyperthermia causative variant as designated by the EMHG Non causative variant Uncertain susceptibility These results do not eliminate the chance that this patient is susceptible to Malignant Hyperthermia. The genetic cause of about half of all MH survivors, with MH susceptibility confirmed by contracture test, remains unknown Clinical findings, family history, further genetic testing and other laboratory data should guide use of halogenated volatile anesthetics or depolarizing muscle relaxants Strong PA166180457
Open in a new tab

Similarly, in CFTR, the metabolizer phenotype may indicate non‐responsiveness to Ivacaftor in Cystic Fibrosis (CF) patients (Table 11). The metabolizer phenotype for this gene is based on the CFTR variant list included in the Ivacaftor label in PharmGKB, as well as the presence of G551D and F508del variants. Other CFTR variants with therapeutic recommendations within the PharmGKB database are also taken into consideration for dosing guidelines. If an individual is heterozygous or homozygous for G551D or any variant included in the CFTR variant list, then they are considered Ivacaftor responsive in CF patients. However, if an individual is homozygous for the F508del variant, they are considered Ivacaftor non‐responsive in CF patients (Whirl‐Carrillo et al., 2021). Additionally, clinically actionable variants in CFTR are also included in the monogenic disease risk and carrier status sections of the genome report.

Table 11.

CFTR Metabolizer Phenotype Classifications

Gene Medication(s) Genotype Metabolizer phenotype Implications Recommendation Strength of recommendation PharmGKB ID
CFTR Ivacaftor

G551D/X

CFTR variant/X

Ivacaftor responsive in CF patients Significant improvement in lung function, weight, risk of pulmonary exacerbation, patient reported outcomes, and reduction in sweat chloride concentrations through enhanced CFTR channel activity (increase probability of open channel) Use ivacaftor according to the product label Strong PA166114461
CFTR Ivacaftor F508del/F508del Ivacaftor non‐responsive in CF patients Ivacaftor treatment is recommended only in cystic fibrosis (CF) patients that are either homozygous or heterozygous for certain CFTR variants. Ivacaftor not recommended Moderate PA166114461
Open in a new tab

In the case of rasburicase and G6PD, the metabolizer phenotype is determined through consideration of variant classes as defined by WHO classification, and the sex of the patient (Table 12). Variants in class I‐III are considered deficient, while class IV is defined as non‐deficient. If a male patient is heterozygous or homozygous for a class IV variant, and a female patient is homozygous for a class IV variant, they are reported as having a Normal Metabolizer phenotype. In contrast, if a male patient is heterozygous or homozygous for a class I‐III variant, and a female patient is homozygous for a class I‐III variant, they are reported as a Deficient metabolizer of rasburicase. Finally, if a female patient has a genotype including one class IV variant and one class I‐III variant, their metabolizer phenotype is reported as Variable. Due to X‐linked mosaicism, female patients with this genotype may display either the deficient or non‐deficient phenotype. However, an enzyme activity test can be used in place to assign G6PD phenotypes in these possible cases.

Table 12.

G6PD Metabolizer Phenotype Classifications

Gene Medication(s) Genotype Metabolizer phenotype Implications Recommendation Strength of recommendation PharmGKB ID
G6PD Rasburicase Male: class IV/X, Female: class IV/class IV Normal metabolizer Low or reduced risk of hemolytic anemia No reason to withhold rasburicase based on G6PD status Strong PA166119846
G6PD Rasburicase Male: class I‐III/X, Female: class I‐III/class I‐III Deficient At risk of acute hemolytic anemia Rasburicase is contraindicated; alternatives include allopurinol Strong PA166119846
G6PD Rasburicase Female: class IV/class I‐III Variable Unknown risk of hemolytic anemia To ascertain that G6PD status is normal, enzyme activity must be measured; alternatives include allopurinol Moderate PA166119846
Open in a new tab

For warfarin dosing (Table 13), the genotypes of CYP2C9, CYP4F2, VKORC1, and the carrier status of rs12777823 contributed to generate formal recommendations for a given patient. For CYP2C9, alleles *2 and *3 were reported as Poor Function phenotypes, while alleles *5, *6, *8, and *11 were reported as Intermediate Metabolizers. If a patient has a CYP2C9 genotype made up of an allele from the Poor Function classification and an allele from the Intermediate Metabolizer classification, both recommendations are included in the main body of the report. In CYP4F2, the rs2108622 T variant was reported as an Increased Metabolizer phenotype, while in VKORC1, the 1639G>A variant was reported as a Poor Function phenotype. Lastly, if a patient was a carrier of the rs12777823 A variant, they will be assigned a Poor Function phenotype, while a rs12777823 T variant is assigned as Unknown for warfarin.

Table 13.

CYP2C9, VKORC1, and CYP4F2 Phenotype Classifications for Warfarin

Gene Medication(s) Variant Metabolizer phenotype Recommendation Strength of recommendation PharmGKB ID
CYP2C9 Warfarin *2 Poor function Based on CYP2C9 *2, administer lower dosage through validated published pharmacogenomics algorithms.

Non‐African Ancestry: Strong

African Ancestry: Moderate

 PA166104949
CYP2C9 Warfarin *3 Poor function Based on CYP2C9 *3, administer lower dosage through validated published pharmacogenomics algorithms.

Non‐African Ancestry: Strong

African Ancestry: Moderate

 PA166104949
CYP2C9 Warfarin *5 Intermediate metabolizer Based on CYP2C9 5*, calculate dosage from validated pharmacogenetic algorithms.

Non‐African Ancestry: Optional

African Ancestry: Moderate

 PA166104949
CYP2C9 Warfarin *6 Intermediate metabolizer Based on CYP2C9 *6, calculate dosage from validated pharmacogenetic algorithms.

Non‐African Ancestry: Optional

African Ancestry: Moderate

 PA166104949
CYP2C9 Warfarin *8 Intermediate metabolizer Based on CYP2C9 *8, calculate dosage from validated pharmacogenetic algorithms.

Non‐African Ancestry: Optional

African Ancestry: Moderate

 PA166104949
CYP2C9 Warfarin *11 Intermediate metabolizer Based on CYP2C9 *11, calculate dosage from validated pharmacogenetic algorithms.

Non‐African Ancestry: Optional

African Ancestry: Moderate

 PA166104949
CYP2C9 Warfarin Normal metabolizer Based on the variants present in CYP2C9, administer recommended dosage. N/A N/A
rsID Warfarin rs12777823 A Poor function Based on rs12777823 carrier A, decrease calculated dosage by 10%‐25%. African Ancestry: Moderate PA166104950
rsID Warfarin Normal metabolizer Based on the absence of rs12777823, administer recommended dosage. N/A N/A
rsID Warfarin rs12777823 T Unknown Based on the rs12777823 carrier T, you have a genetic change that may alter how effective this medication is for you. However, at this time we are not entirely sure what this change means or how/if this alters standard dosing recommendations. Changes to current medication(s) or dosing are not recommended based on this finding.
CYP4F2 Warfarin rs2108622 T Increased function Based on rs2108622 carrier T in CYP4F2, increase dose by 5%‐10%. Non‐African Ancestry: Optional PA166104949
CYP4F2 Warfarin Normal metabolizer Based on rs2108622 in CYP4F2, administer recommended dosage. N/A N/A
VKORC1 Warfarin 1639G>A Poor function Based on VKORC1 1639G>A, administer lower dosage through validated published pharmacogenomics algorithms.

Non‐African Ancestry: Strong

African Ancestry: Moderate

 PA166104949
VKORC1 Warfarin Normal metabolizer Based on the absence of VKORC1 1639G>A, administer recommended dosage. N/A N/A
Open in a new tab

For each of the four regions affecting warfarin, if a patient was not a carrier of the variants listed with formal recommendations from PharmGKB, they are reported to have a Normal Metabolizer phenotype. The Normal Metabolizer assignments for these four regions are performed in consideration of the CPIC guidelines for pharmacogenetics‐guided warfarin dosing (Johnson et al., 2017). Additionally, the self‐reported ancestry of the patient also has considerations for warfarin dosing, and is included in the strength of recommendations on the report (Whirl‐Carrillo et al., 2021).

For the genes CYP2C9, DPYD, and CYP2D6, the activity score was also taken into consideration for a combination of certain medications and metabolizer phenotypes (Table 14). An activity score is defined as the sum of values assigned to each allele of a given gene, and typically ranges between 0 and 3.0, although some genes such as CYP2D6 have activity scores that can exceed this range (Crews et al., 2014). For the case of CYP2D6, in which the copy number of variants can affect recommendations, a score of 0 indicates a poor metabolizer status, 0.25 to 1.0 indicates an intermediate metabolizer status, 1.25 to 2.25 indicates a normal metabolizer status, and >2.25 indicates an ultrarapid or extensive metabolizer status. In the cases of CYP2C9 and DPYD, 0 to 0.5 indicates a poor metabolizer status, 1.0 to 1.5 indicates an intermediate metabolizer status, and 2.0 indicates a normal metabolizer status (Whirl‐Carrillo et al., 2021). The activity scores for these genes are appended from the PharmGKB diplotype‐phenotype tables and used to assign formal recommendations that may differ slightly between score assignments within the same metabolizer phenotype classification. For example, in Table 14, the implications and recommendations for CYP2C9 genotypes at an activity score of 1 and 1.5 vary, despite being both assigned as intermediate metabolizers.

Table 14.

Pharmacogenomic Activity Scores for CYP2C9, DPYD, CYP2D6

Gene Medication(s) Activity score Metabolizer phenotype Implications Recommendation Strength of recommendation PharmGKB ID
CYP2C9

Celecoxib

Flurbiprofen

Ibuprofen

Lornoxicam

 1.5 Intermediate metabolizer Mildly reduced metabolism. Initiate therapy with recommended starting dose. In accordance with the prescribing information, use the lowest effective dosage for shortest duration consistent with individual patient treatment goals. Moderate PA166191841
CYP2C9

Celecoxib

Flurbiprofen

Ibuprofen

Lornoxicam

 1 Intermediate metabolizer Moderately reduced metabolism; higher plasma concentrations may increase probability of toxicities. Initiate therapy with lowest recommended starting dose. Titrate dose upward to clinical effect or maximum recommended dose with caution. In accordance with the prescribing information, use the lowest effective dosage for shortest duration consistent with individual patient treatment goals. Carefully monitor adverse events such as blood pressure and kidney function during course of therapy. Moderate PA166191841
CYP2C9

Fosphenytoin

Phenytoin

 1.5 Intermediate metabolizer Slightly reduced phenytoin metabolism; however, this does not appear to translate into increased side effects. No adjustments needed from typical dosing strategies. Subsequent doses should be adjusted according to therapeutic drug monitoring, response, and side effects. An HLA‐B*15:02 negative test does not eliminate the risk of phenytoin‐induced SJS/TEN, and patients should be carefully monitored according to standard practice. Moderate PA166122806
CYP2C9

Fosphenytoin

Phenytoin

 1 Intermediate metabolizer Reduced phenytoin metabolism, higher plasma concentrations will increase probability of toxicities. For first dose, use typical initial or loading dose. For subsequent doses, use approximately 25% less than typical maintenance dose. Subsequent doses should be adjusted according to therapeutic drug monitoring, response and side effects. An HLA‐B*15:02 negative test does not eliminate the risk of phenytoin‐induced SJS/TEN, and patients should be carefully monitored according to standard practice. Moderate PA166122806
CYP2C9 Meloxicam 1.5 Intermediate metabolizer Mildly reduced metabolism. Initiate therapy with recommended starting dose. In accordance with the prescribing information, use the lowest effective dosage for shortest duration consistent with individual patient treatment goals. Moderate PA166192301
CYP2C9 Meloxicam 1 Intermediate metabolizer Moderately reduced metabolism; higher plasma concentrations may increase probability of toxicities. Initiate therapy with 50% of the lowest recommended starting dose. Titrate dose upward to clinical effect or 50% of the maximum recommended dose with caution. In accordance with the meloxicam prescribing information, use the lowest effective dosage for shortest duration consistent with individual patient treatment goals. Upward dose titration should not occur until after steady state is reached (at least 7 days). Carefully monitor adverse events such as blood pressure and kidney function during course of therapy. Alternatively, consider alternative therapy. Choose an alternative therapy not metabolized by CYP2C9 or not significantly impacted by CYP2C9 genetic variants in vivo or choose an NSAID metabolized by CYP2C9 but with a shorter half‐life. Moderate PA166192301
CYP2C9 Piroxicam 1.5 Intermediate metabolizer Mildly reduced metabolism. Initiate therapy with recommended starting dose. In accordance with the prescribing information, use the lowest effective dosage for shortest duration consistent with individual patient treatment goals. Moderate PA166192321
CYP2C9 Piroxicam 1 Intermediate metabolizer Moderately reduced metabolism; higher plasma concentrations may increase probability of toxicities. Choose an alternative therapy not metabolized by CYP2C9 or not significantly impacted by CYP2C9 genetic variants in vivo or choose an NSAID metabolized by CYP2C9 but with a shorter half‐life. Moderate PA166192321
CYP2C9 Tenoxicam 1.5 Intermediate metabolizer Mildly reduced metabolism. Initiate therapy with recommended starting dose. In accordance with the prescribing information, use the lowest effective dosage for shortest duration consistent with individual patient treatment goals. Moderate PA166192341
CYP2C9 Tenoxicam 1 Intermediate metabolizer Moderately reduced metabolism; higher plasma concentrations may increase probability of toxicities. Choose an alternative therapy not metabolized by CYP2C9 or not significantly impacted by CYP2C9 genetic variants in vivo or choose an NSAID metabolized by CYP2C9 but with a shorter half‐life. Optional PA166192341
DPYD Capecitabine 1.5 Intermediate metabolizer Decreased DPD activity (leukocyte DPD activity at 30% to 70% that of the normal population) and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Reduce starting dose by 50% followed by titration of dose based on toxicity or therapeutic drug monitoring (if available). Patients with the c.[2846A>T];[2846A>T] genotype may require >50% reduction in starting dose. Moderate PA166109594
DPYD Capecitabine 1 Intermediate metabolizer Decreased DPD activity (leukocyte DPD activity at 30% to 70% that of the normal population) and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Reduce starting dose by 50% followed by titration of dose based on toxicity or therapeutic drug monitoring (if available). Patients with the c.[2846A>T];[2846A>T] genotype may require >50% reduction in starting dose. Strong PA166109594
DPYD Capecitabine 0.5 Poor metabolizer Complete DPD deficiency and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Avoid use of 5‐ fluorouracil or 5‐fluorouracil prodrug‐based regimens. In the event, based on clinical advice, alternative agents are not considered a suitable therapeutic option, 5‐fluorouracil should be administered at a strongly reduced dose with early therapeutic drug monitoring. Strong PA166109594
DPYD Capecitabine 0 Poor metabolizer Complete DPD deficiency and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Avoid use of 5‐fluorouracil or 5‐fluorouracil prodrug‐based regimens. Strong PA166109594
DPYD Fluorouracil 1.5 Intermediate metabolizer Decreased DPD activity (leukocyte DPD activity at 30% to 70% that of the normal population) and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Reduce starting dose by 50% followed by titration of dose based on toxicity or therapeutic drug monitoring (if available). Patients with the c.[2846A>T];[2846A>T] genotype may require >50% reduction in starting dose. Moderate PA166122686
DPYD Fluorouracil 1 Intermediate metabolizer Decreased DPD activity (leukocyte DPD activity at 30% to 70% that of the normal population) and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Reduce starting dose by 50% followed by titration of dose based on toxicity or therapeutic drug monitoring (if available). Patients with the c.[2846A>T];[2846A>T] genotype may require >50% reduction in starting dose. Strong PA166122686
DPYD Fluorouracil 0.5 Poor Metabolizer Complete DPD deficiency and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Avoid use of 5‐ fluorouracil or 5‐fluorouracil prodrug‐based regimens. In the event, based on clinical advice, alternative agents are not considered a suitable therapeutic option, 5‐fluorouracil should be administered at a strongly reduced dose with early therapeutic drug monitoring. Strong PA166122686
DPYD Fluorouracil 0 Poor metabolizer Complete DPD deficiency and increased risk for severe or even fatal drug toxicity when treated with fluoropyrimidine drugs. Avoid use of 5‐fluorouracil or 5‐fluorouracil prodrug‐based regimens. Strong PA166122686
CYP2D6 Atomoxetine 1 Intermediate metabolizer Possibly higher atomoxetine concentrations as compared to normal metabolizers but questionable clinical significance. Normal metabolizers may be at an increased risk of increased discontinuation as compared to poor metabolizers. Initiate with a dose of 0.5 mg/kg and increase to 1.2 mg/kg/day after 3 days If no clinical response and in the absence of adverse events after 2 weeks, consider obtaining a peak plasma concentration (1 to 2 hr after dose administered). If <200 ng/ml, consider a proportional increase in dose to approach 400 ng/ml. Moderate PA166181885
CYP2D6 Atomoxetine 1 Intermediate metabolizer Decreased metabolism of atomoxetine and higher atomoxetine concentrations as compared to normal metabolizers. Individuals with activity score of 1.0 with CYP2D6*10 may be at an increased risk of increased discontinuation as compared to poor metabolizers. Initiate with a dose of 0.5 mg/kg/day and if no clinical response and in the absence of adverse events after 2 weeks, consider obtaining a plasma concentration 2‐4 hr after dosing. If response is inadequate and concentration is <200 ng/ml, consider a proportional dose increase to achieve a concentration to approach 400 ng/ml. If unacceptable side effects are present at any time, consider a reduction in dose. Moderate PA166181885
CYP2D6 Atomoxetine 0.5 Intermediate metabolizer Decreased metabolism of atomoxetine and higher atomoxetine concentrations as compared to normal metabolizers. Intermediate metabolizers may be at an increased risk of discontinuation as compared to poor metabolizers. Initiate with a dose of 0.5 mg/kg/day and if no clinical response and in the absence of adverse events after 2 weeks, consider obtaining a plasma concentration 2‐4 hr after dosing. If response is inadequate and concentration is <200 ng/ml, consider a proportional dose increase to achieve a concentration to approach 400 ng/ml. If unacceptable side effects are present at any time, consider a reduction in dose. Moderate PA166181885
CYP2D6 Tamoxifen 1 Intermediate metabolizer Lower endoxifen concentrations compared to normal metabolizers; higher risk of breast cancer recurrence, event‐free and recurrence‐free survival compared to normal metabolizers. Consider hormonal therapy such as an aromatase inhibitor for postmenopausal women or aromatase inhibitor along with ovarian function suppression in premenopausal women, given that these approaches are superior to tamoxifen regardless of CYP2D6 genotype. If aromatase inhibitor use is contraindicated, consideration should be given to use a higher but FDA approved tamoxifen dose (40 mg/day). Avoid CYP2D6 strong to weak inhibitors. Optional PA166176068
CYP2D6 Tamoxifen 1 Intermediate metabolizer Lower endoxifen concentrations compared to normal metabolizers; higher risk of breast cancer recurrence, event‐free and recurrence‐free survival compared to normal metabolizers. Consider hormonal therapy such as an aromatase inhibitor for postmenopausal women or aromatase inhibitor along with ovarian function suppression in premenopausal women, given that these approaches are superior to tamoxifen regardless of CYP2D6 genotype. If aromatase inhibitor use is contraindicated, consideration should be given to use a higher but FDA approved tamoxifen dose (40 mg/day). Avoid CYP2D6 strong to weak inhibitors. Moderate PA166176068
CYP2D6 Tamoxifen 0.5 Intermediate metabolizer Lower endoxifen concentrations compared to normal metabolizers; higher risk of breast cancer recurrence, event‐free and recurrence‐free survival compared to normal metabolizers. Consider hormonal therapy such as an aromatase inhibitor for postmenopausal women or aromatase inhibitor along with ovarian function suppression in premenopausal women, given that these approaches are superior to tamoxifen regardless of CYP2D6 genotype. If aromatase inhibitor use is contraindicated, consideration should be given to use a higher but FDA approved tamoxifen dose (40 mg/day). Avoid CYP2D6 strong to weak inhibitors. Moderate PA166176068
Open in a new tab

For HLA‐A and HLA‐B genotypes determined using Basic Protocol 1, the listed variants for each gene are also used to determine their metabolizer phenotypes (Table 15). For HLA‐B, a metabolizer phenotype of Increased Risk is assigned with the presence of the *57:01 variant for Abacavir, the presence of *58:01 for Allopurinol, and the presence of the *15:01 variant for Carbamazepine, Fosphenytoin, Phenytoin, and Oxcarbazepine. For HLA‐A, a metabolizer phenotype of Increased Risk is assigned with the presence of *31:01 for Carbamazepine. Otherwise, in both HLA‐A and HLA‐B, they are assigned the metabolizer phenotype of Normal or Reduced Risk.

Table 15.

HLA‐A, HLA‐B Metabolizer Phenotype Table

Gene Medication(s) Variant Metabolizer phenotype Implications Recommendation Strength of recommendation PharmGKB ID
HLA‐A Carbamazepine Normal or reduced risk Normal or reduced risk of carbamazepine‐induced SJS/TEN, DRESS and MPE Use carbamazepine per standard dosing guidelines. Strong PA166105008
HLA‐A Carbamazepine *31:01 Increased risk Increased risk of carbamazepine‐induced SJS/TEN, DRESS and MPE If patient is carbamazepine‐naive and alternative agents are available, do not use carbamazepine. Strong PA166105008
HLA‐B Abacavir Normal or reduced risk Low or reduced risk of abacavir hypersensitivity Use abacavir per standard dosing guidelines. Strong PA166104997
HLA‐B Abacavir *57:01 Increased risk Significantly increased risk of abacavir hypersensitivity Abacavir not recommended. Strong PA166104997
HLA‐B Allopurinol Normal or reduced risk Low or reduced risk of allopurinol SCAR Use allopurinol per standard dosing guidelines. Strong PA166105003
HLA‐B Allopurinol *58:01 Increased Risk Significantly increased risk of allopurinol SCAR Allopurinol is contraindicated. Strong PA166105003
HLA‐B Carbamazepine Normal or Reduced Risk Normal or reduced risk of carbamazepine‐induced SJS/TEN, DRESS and MPE. Use carbamazepine per standard dosing guidelines. Strong PA166105008
HLA‐B Carbamazepine *15:02 Increased risk Increased risk of carbamazepine‐induced SJS/TEN If patient is carbamazepine‐naive, do not use carbamazepine. Strong PA166105008
HLA‐B

Fosphenytoin

Phenytoin

 Normal or reduced risk Normal phenytoin metabolism No adjustments needed from typical dosing strategies. Subsequent doses should be adjusted according to therapeutic drug monitoring, response, and side effects. An HLA‐B*15:02 negative test does not eliminate the risk of phenytoin‐induced SJS/TEN, and patients should be carefully monitored according to standard practice. Strong PA166122806
HLA‐B

Fosphenytoin

Phenytoin

 *15:02 Increased risk Increased risk of phenytoin‐induced SJS/TEN

If patient is phenytoin‐naive, do not use phenytoin/fosphenytoin. Avoid carbamazepine and oxcarbazepine.

Optional recommendation: If the patient has previously used phenytoin continuously for longer than three months without incidence of cutaneous adverse reactions, cautiously consider use of phenytoin in the future. The latency period for drug‐induced SJS/TEN is short with continuous dosing and adherence to therapy (4‐28 days), and cases usually occur within three months of dosing.

 Strong PA166122806
HLA‐B Oxcarbazepine Normal or reduced risk Normal or reduced risk of oxcarbazepine‐induced SJS/TEN Use oxcarbazepine per standard dosing guidelines. Strong PA166176623
HLA‐B Oxcarbazepine *15:02 Increased risk Increased risk of oxcarbazepine‐induced SJS/TEN If patient is oxcarbazepine‐naive, do not use oxcarbazepine. Strong PA166176623
Open in a new tab

For IFNL3, the genotype at rs12979860 is used to determine the metabolizer phenotype, with the genotype CC indicating a Favorable Response, while CT or TT indicate an Unfavorable Response (Table 16). Otherwise, a metabolizer phenotype of Unknown is assigned for IFNL3.

Table 16.

IFNL3 Metabolizer Phenotype Table

Gene Medication(s) Variant Metabolizer phenotype Implications Recommendation Strength of recommendation PharmGKB ID
IFNL3

Peginterferon alfa‐2a

Peginterferon alfa‐2b

Ribavirin

 rs12979860 CC Favorable response

For PEG‐IFN ALPHA and RBV based on the variation rs12979860

Approximately 70% chance for SVR after 48 weeks of treatment. Consider implications before initiating PEG‐IFN alpha and RBV containing regimens

For protease inhibitor combinations with PEG‐IFN ALPHA and RBV based on the variation rs12979860

Approximately 90% chance for SVR after 24‐48 weeks of treatment. Approximately 80%‐90% of patients are eligible for shortened therapy (24‐28 weeks vs. 48 weeks). Weighs in favor of using PEG‐IFN alpha and RBV containing regimen

 Strong PA166110235
IFNL3

Peginterferon alfa‐2a

Peginterferon alfa‐2b

Ribavirin

 rs12979860 CT/TT Unfavorable Response

For PEG‐IFN ALPHA and RBV based on the variation rs12979860

Approximately 30% chance for SVR after 48 weeks of treatment. Consider implications before initiating PEG‐IFN alpha and RBV containing regimens

For protease inhibitor combinations with PEG‐IFN ALPHA and RBV based on the variation rs12979860

Approximately 60% chance for SVR after 24‐48 weeks of treatment. Approximately 50% of patients are eligible for shortened therapy (24‐28 weeks). Consider implications before initiating PEG‐IFN and RBV containing regimens. For PEG‐IFN ALPHA and RBV: Approximately 70% chance for SVR c after 48 weeks of treatment. Consider implications before initiating PEG‐IFN alpha and RBV containing regimens.

 Strong PA166110235
Open in a new tab

Sample_PHARMACOGENOMICS_SUPPLEMENTARY/ folder

The PHARMACOGNEOMICS_SUPPLEMENTARY folder contains supplementary data and the intermediary files generated in the Pharmacogenomics pipeline. The final_stargazer_output.txt file contains concatenated results from the Stargazer software for all the pharmacogenes analyzed in this pipeline. The rsID_Warfarin.txt file contains the results from the bcftools view command for rsid12777823 to determine an individual's warfarin metabolizer status. Lastly, the Target_GDF_RYR1.gdf file is also included in this folder, for the purposes of analyzing the possible presence of structural variants in CYP2D6 with the Stargazer software. The .gdf file is generated using the GATK software, and analyzed using the control gene RYR1 as recommended from Lee, Wheeler, Thummel, and Nickerson (2019). The VDR gene .bed file is also included in the Pharmacogenomics Configuration folder to be used as an alternative control gene, and was also outlined in (Lee et al., 2019).

COMMENTARY

Background Information

Basic Protocol 1: HLA genotyping

HLA stands for “Human Leukocyte Antigen,” and is associated with the body's regulation and the function of immune response. An individual's HLA type is grouped into a different class depending on the genes that are present. For example, class I consists of the HLA‐A, HLA‐B, and HLA‐C genes, and class II consists of the HLA‐DRB1 and HLA‐DQB1 genes (Matzaraki, Kumar, Wijmenga, & Zhernakova, 2017). Studies have shown that certain genetic changes within the HLA region may be associated with susceptibility to and severity of SARS‐CoV‐2 infection. For example, studies have suggested that having a certain genetic variant in class I HLA‐B, HLA‐B*46:01 may result in an individual being more susceptible to a SARS‐CoV‐2 infection. Conversely, having a different genetic variant in class I HLA‐B, HLA‐B*15:03 may result in having greater protection against a SARS‐CoV‐2 infection (Anastassopoulou, Gkizarioti, Patrinos, & Tsakris, 2020). Certain genetic variants in the HLA region have previously been found to be associated with an increased risk for autoimmune conditions such as type 1 diabetes, ankylosing spondylitis, and rheumatoid arthritis (Gao et al., 2019). Therefore, analysis of HLA status may provide valuable information on potential disease risk in addition to providing novel information on COVID‐19 immune response. As we acquire more information about the association between HLA, autoimmune conditions, and COVID‐19, knowing HLA status from a genome report may help inform response to SARS‐CoV‐2, vaccine response, early candidates for vaccine development, treatment, and clinical decision‐making tools for individuals with a past infection.

The HLA typing pipeline was developed to estimate the most probable HLA genotypes for a given individual and indicate potential disease risks associated with certain HLA alleles. The software HLA‐VBSeq v2 (Mimori et al., 2019) was used to estimate the most probable HLA types from whole‐genome‐sequencing patient data. Read names were extracted from BAM files for 22 HLA loci, including HLA‐A, ‐B, ‐C. ‐DQA1, ‐DQB1, ‐DPB1, and ‐DRB1. The extracted reads were then aligned to the HLA v2 database from the HLA‐VBSeq package, which is based on the IMGT/HLA database Release 3.31.0 and the Japanese HLA reference dataset (Mimori et al., 2019). The HLA types were then estimated for paired‐end read data using the HLA‐VBSeq package along with the average depth of coverage for each HLA allele. The top called allele for each HLA loci was reported to four‐digit resolution. The top called alleles for the seven major HLA loci were then compared to an in‐house database containing currently known HLA‐disease associations as reported from the literature. An increased disease risk was reported if a patient's HLA genotype or genotypes were found to be associated with a relevant disease or condition. Alleles were only reported if they met HLA‐VBSeq's default coverage threshold for HLA typing.

HLA alleles associated with increased autoimmune disease risk were identified through a review of literature available on PubMed (https://pubmed.ncbi.nlm.nih.gov). Papers reviewing the association between the HLA locus and human autoimmune disease in general were identified using the following search terms: “HLA” OR “MHC” AND “human disease” OR “autoimmune disease”. A list of autoimmune diseases consistently reported to be associated with HLA was compiled as follows: Type 1 Diabetes (T1D), Celiac Disease (CD), Rheumatoid Arthritis (RA), Ankylosing Spondylitis (AS), Behçets Disease (BD), Multiple Sclerosis (MS), and Graves’ Disease (GD). Subsequent searches were conducted on PubMed for review articles highlighting the role of HLA in each of the above diseases. The following search terms were used: “Type 1 Diabetes” OR “T1D” OR “Insulin Dependent Diabetes Mellitus” OR “IDDM” OR “Celiac Disease” OR “Coeliac Disease” OR “Rheumatoid Arthritis” OR “Ankylosing Spondylitis” OR “Behçet's Disease” OR “Behçet's Syndrome” OR “Multiple Sclerosis” OR “Graves’ Disease” AND “HLA” OR “MHC” AND “association”OR “risk”. From this search, a list of HLA alleles consistently reported to be associated with each disease was compiled, followed by an evaluation of the significance of disease risk. Subsequent searches were conducted to identify the disease risks associated with each HLA allele. The following search terms were used. For T1D: “Type 1 Diabetes”, “T1D”, “IDDM”, “Insulin‐Dependent Diabetes Mellitus”, “HLA”, “MHC”, “DRB1”, “DQA1”, “DQB1”, “DR”, “class II”, “allele”, “HLA genotype”, “risk”, “disease risk”, “genetic risk”, “case‐control”. For CD: “Celiac Disease”, “Coeliac Disease”, “HLA”, “MHC”, “HLA genotype”, “disease risk”, “risk”, “genetic risk”, “HLA‐DQ”, “case‐control”. For RA: “Rheumatoid Arthritis”, “HLA”, “MHC”, “association”, “risk”, “genetic risk”, “HLA‐DRB1”, “shared epitope”, “susceptibility”, “case‐control”. For AS: “ankylosing spondylitis”, “HLA”, “HLA‐B27”, “susceptibility”, “association”, “case‐control”. For BD: “Behçet's disease”, “Behçet Disease”, “HLA”, “MHC”, “HLA‐B*51”, “HLA‐B51”, “association”, “class I”, “case‐control”. For MS: “Multiple Sclerosis”, “HLA”, “genetics”, “association”, “susceptibility”, “HLA‐DRB1*15”, “HLA‐DR15”. For GD: “Graves’ Disease”, “HLA”, “class II”, “DQA1*0501”, “case‐control”, “association”. Papers were included if they were a case‐control study, or a meta‐analysis of a case‐control study, and were published after 1990. Disease risks were included if they were published in a case‐control study or a meta‐analysis of case‐control studies, were reported as an odds ratio (OR), or relative risk (RR), had an OR RR > 1, and were found to be statistically significant by the published study. The full list of references used for the development of the in‐house HLA disease association database can be found in File S6 in Supporting Information.

Basic Protocol 2: Ancestry estimation

A person's race typically refers to their physical appearance and characteristics, such as their skin and eye color. A person's ethnicity refers to communality in cultural heritage, language, social practice, traditions, and geopolitical factors. Genetic ancestry is often inferred using ancestry informative markers (AIMs) based on genetic/genomic data, and can thus be considered a biological variable. Genetic ancestry can sometimes disagree with self‐reported race and/or ethnicity (Mersha & Abebe, 2015). Identifying genetic ancestry, currently used primarily in the recreational genetic testing context, may be important for understanding which populations are disproportionately affected by COVID‐19, as well as if there are any significant differences in COVID‐19 susceptibility and response across individuals of various genetic ancestries. To illustrate current understanding of COVID‐19 disease outcomes among various populations, the American Center for Disease Control (CDC) shared statistics in March 2021 that highlight COVID‐19 infection rates in the U.S.A. as categorized by race and ethnicity. The CDC identified that American Indian, Non‐Hispanic persons had a 1.6× increase in risk of infection compared to White, Non‐Hispanic persons, a 3.5× increase in hospitalization, and a 2.4× increase in death. Similarly for Asian, Black, and Hispanic persons, a 0.7×, 1.1×, and 2.0× increase in risk of infection were reported, respectively, a 1.0×, 2.8×, and 3.0× increased risk of hospitalization, and a 1.0×, 1.9×, and a 2.3× risk of death compared to White, Non‐Hispanic persons (COVID‐NET, accessed April 23rd, 2021; see Internet Resources). These statistics note the highly elevated rates in cases, hospitalizations, and death across various populations compared to White, Non‐Hispanic Americans, particularly in American‐Indigenous, Black, and Hispanic persons. Alongside these findings, other studies conducted across the U.S.A. and U.K. have also investigated COVID‐19 risk among diverse populations, and have concluded that patients of Black, Asian, and Hispanic ancestry were more likely to be infected with COVID‐19, as well as being at a higher risk both for admission to intensive care and for death (Sze et al., 2020). In addition, a genetic association study identified a particular gene cluster on chromosome 3 (45,859,651‐45,909,024, hg19) from Neanderthals that may indicate risk for respiratory failure after infection with COVID‐19, as well as act as a risk factor for other severe symptoms and hospitalization. It is carried by ∼50% of people in South Asia and ∼16% of people in Europe, suggesting another potential link between COVID‐19 risk and genetic ancestry in some infected individuals (Zeberg & Pääbo, 2020). Although there are currently major limitations when using ancestry‐based algorithms to predict genetic ancestry, we sought to incorporate this information into a comprehensive report for the purpose of evaluating the utility of genomic results as part of the GENCOV study (Taher et al., 2021).

Large‐scale continental individual ancestry was estimated through the combination of patient genotypes with those from the reference dataset, HapMap3 (The International HapMap 3 Consortium et al., 2010). The International HapMap 3 Consortium have genotyped 1.6 million common single nucleotide polymorphisms (SNPs) in 1184 reference individuals from 11 global populations (The International HapMap 3 Consortium et al., 2010, see Table 17). Various filtering strategies were applied to both the patient genotypes and the reference dataset, removing duplicate SNPs and correcting for chromosome and position mismatches using bash scripting and PLINK software. Furthermore, the genetic variants used for estimation were pruned for variants in linkage disequilibrium (LD) with an r2 > 0.2 in a 50 kb window. Variants were also excluded if they fell within genomic ranges of known high‐LD structure following the PlinkQC protocol for ancestry estimation using the HapMap3 reference dataset (Meyer, 2021a, Meyer, 2021b). Ancestry estimation can alternatively be performed using reference data available through The 1000 Genomes Project initiative to increase the accuracy of estimation reporting and for validation purposes (1000 Genomes Project Consortium et al., 2015, Meyer, 2021b).

Table 17.

HapMap3 Reference Dataset

Population Population abbreviation Number of samples
African ancestry in Southwest USA ASW 83
Utah residents with Northern and Western European ancestry from the CEPH collection CEU 165
Han Chinese in Beijing, China CHB 84
Chinese in Metropolitan Denver, Colorado CHD 85
Gujarati Indians in Houston, Texas GIH 88
Japanese in Tokyo, Japan JPT 86
Luhya in Webuye, Kenya LWK 90
Maasai in Kinyawa, Kenya MKK 171
Mexican ancestry in Los Angeles, California MXL 77
Toscani in Italia TSI 88
Yoruba in Ibadan, Nigeria YRI 167
TOTAL 1184
Open in a new tab

ADMIXTURE was used to output estimated patient ancestry against the reference dataset following filtration. ADMIXTURE is a model‐based ancestry estimation software, modeling the probability of the observed genotypes using ancestry proportions and population allele frequencies (Alexander et al., 2009). In‐house scripts were then used to generate a visual summary of patient admixed ancestry.

Basic Protocol 3: Pharmacogenomics

Pharmacogenomics is the study of relationships between genetic variation and its influence on how bodies respond to medications. More often than not, prescribed medications are assumed to be equally effective in all patients, regardless of factors such as age, body mass index, sex, or diet, provided that they are given appropriately adjusted doses. However, genetic variation among patients may lead to varying drug responses even when accounting for most of these traits. Genetic variation can also alter drug metabolism by affecting how much of a given medication is absorbed into the body and how efficiently it is able to travel to its intended destination, as well as how quickly it gets broken down. If a given individual's genetic variation alters any of these processes, certain medications may not work as well or cause side effects. Pharmacogenomic testing provides valuable insight into how individuals with particular genetic variants might respond differently to medication(s). Therefore, identifying gene variants with relevant pharmacogenomic variants can allow healthcare providers to give more personalized recommendations related to drug dosage and/or treatment regimens.

The pharmacogenomics reports were based on the recommendations from The Pharmacogenomics Knowledge Base (Whirl‐Carrillo et al., 2021, accessed as of October 2021). PharmGKB is a resource containing clinically actionable gene‐drug associations and genotype‐phenotype relationships (Thorn, Klein, & Altman, 2013). From PharmGKB, dosing guidelines were taken from the Clinical Pharmacogenetics Implementation Consortium (CPIC) for CPIC Level A and PharmGKB Level of Evidence 1A. Stargazer is a software application that detects single nucleotide variants, insertion‐deletion variants, and structural variants, including gene deletions, duplications, and conversions, for identifying genotypes in pharmacogenes (Lee et al., 2019). Using Stargazer, the genotypes of 17 pharmacogenes were identified, including structural variant analysis in CYP2D6. Additionally, the genotypes of two HLA loci, HLA‐A and HLA‐B, and rs12777823 were also included. The resulting genotypes were then linked to a metabolizer phenotype reference database compiled in‐house. Pharmacogenes that were unable to be genotyped for certain participants through the Stargazer software were included as part of the report Appendix with haplotype candidates listed instead. The metabolizer phenotype database was generated from the gene diplotype‐phenotype tables and the gene‐medication dosing guidelines from PharmGKB and CPIC, accessed as of December 2021. If a genotype was not defined with CPIC Level A and PharmGKB Level of Evidence 1A, a metabolizer phenotype of Unknown and a standardized recommendation were assigned to the given gene medication.

Additional protocols

In addition to the basic protocols outlined above, comprehensive gene panels were developed to aid in variant filtration and assessment. A list of medically actionable gene panels was developed according to specifications of medically actionable genes as reported by the American College of Medical Genetics and Genomics (Miller et al., 2021), The Clinical Genome Resource (ClinGen, Rehm et al., 2015), and Reble et al. (2021) for a comprehensive panel including 208 genes. Additional panels containing genes associated with carrier status and Mendelian, rare genetic diseases were included under the following categories: audiology and ophthalmology, 1387 genes; biochemical and mitochondrial, 944 genes; blood and immunology, 898 genes; cardiovascular, 1000 genes; carrier screening, 920; dermatological, 582 genes; endocrine, 460 genes; gastrointestinal, 612 genes; musculoskeletal, 1619 genes; neurological, 2639 genes; oncology, 735 genes; renal, 860 genes. These comprehensive disease gene panels were compiled using ClinGen, Clinical Genomics Database (CGD, Solomon, Nguyen, Bear, & Wolfsberg, 2013), Online Mendelian Inheritance in Man (OMIM, Hamosh, Scott, Amberger, Valle, & McKusick, 2000), BabySeq (Holm et al., 2018), and The Gene Curation Coalition (GenCC) (DiStefano et al., 2022) gene databases. Genes from these databases were cross‐referenced with panels from reputable genetic testing companies. Additional genes associated with COVID‐19 were identified through a literature review. Periodic updates of this database are needed to incorporate new evidence for gene‐disease relationships as they are identified. The comprehensive gene panel database can be found in File S5 in Supporting Information.

Following the development of gene panels, participant data are processed and analyzed using the Franklin Genoox Platform (https://www.genoox.com) in combination with GATK best practices for variant calling. Reads are aligned using BWA against the GRCh37 human reference, with duplicate reads removed. Variant calling is performed using GATK (version 4.1) and FreeBayes (version 1.1.0). Base calling is performed using bcl2fastq 2.20 (for HiSeq 2500) or HiSeq Analysis Software (for HiSeq X). Reads are mapped to the b37 reference sequence using the bwa‐mem algorithm, and duplicate reads are marked using Picard Tools. Local realignment and base quality score recalibration are also performed using GATK. Variants are called using GATK HaplotypeCaller, followed by variant quality score recalibration (VQSR) for filtration. Variants are identified for further analysis using the following filtering parameters: (1) variants classified as pathogenic or likely pathogenic in ClinVar (Landrum et al., 2018); (2) variants with a Genoox classification of pathogenic/likely pathogenic/leaning pathogenic, aggregated allele frequency ≤5%, ± identified in ClinVar, with an established gene‐disease relationship; (3) variants with a prior classification within Genoox; (4) common risk factor variants (aggregated allele frequency >5%); and (5) rare risk factor variants (aggregated allele frequency ≤5%). Variants determined to be likely pathogenic or pathogenic using ACMG criteria are reported back to participants in a screening context in the report results sections as ‘you are at risk for’ or ‘you are a carrier of’.

Copy number variants (CNVs) are initially detected and annotated using an Annovar‐based custom pipeline developed by The Center for Applied Genomics (TCAG), and were further filtered for variant inclusion based on (1) OMIM morbid genes; (2) ISCA triplo‐ or haplo‐sensitivity scores; (3) size (>20Kb); and, (4) frequency (<1%). CNVs with >90% overlap with known benign, polymorphic CNV regions excluded. Variants are filtered further by quality parameters (e.g., confidence of call), internal and population frequencies, published evidence, ClinVar submissions, validity of gene‐disease relationships, and participant preferences. Example variant interpretation reports can be found in File S1 in Supporting Information.

Two other analyses were implemented in our genome reports according to existing protocols developed by collaborators within the GENCOV study. These include polygenic risk score calculation for six diseases (atrial fibrillation, coronary artery disease, type 2 diabetes, prostate cancer, colorectal cancer, breast cancer; Hao et al., 2022) and ABO, Rh, and other blood group genotyping (Halls et al., 2020, Lane et al., 2018).

A polygenic risk score (PRS) takes into account multiple genetic changes to provide a personalized risk for developing a specific disease. A PRS provides a “relative” risk of developing a disease compared to individuals of similar age, ethnicity, and genetic background. This score does not account for factors other than genetics that may influence an individual's risk for common conditions, including lifestyle factors (e.g., diet and exercise) and family history. Using whole genome sequencing data, polygenic risk scores (PRS) were calculated using existing algorithms developed in Hao et al. (2022). All relevant code for PRS calculations is available and adapted from the following public github repository: https://github.com/MGB‐Personalized‐Medicine/PRS‐adjustment. The scores were developed from GWAS and replicated in multiple cohorts. Each individual carries 0, 1, or 2 risk alleles at each locus, which are weighted by their per‐allele effect size and summed. Unadjusted, raw PRS are calculated with PLINK software and known SNPs and associated effect size (from the discovery GWAS) for type 2 diabetes (6,917,436 SNPs), atrial fibrillation (6,730,541 SNPs), coronary artery disease (6,630,150 SNPs; Khera et al., 2018), female breast cancer (3820 SNPs; Mavaddat et al., 2015, Michailidou et al., 2017), colorectal (81 SNPs; Huyghe et al., 2019), and prostate cancer (147 SNPs; Schumacher et al., 2018). Raw scores are adjusted for ancestry using ancestry‐informative principal components to calculate residualized PRS. Individuals with PRS exceeding a threshold corresponding with an odds ratio (OR) >2 are considered at “increased polygenic risk.” All other PRS results are considered “average polygenic risk.” For example, if a person's OR for developing colon cancer given that they have a certain genetic change(s)/variant(s) is equal to 2, then the individual has a 2‐fold increased risk of developing colorectal cancer compared to individuals without the same genetic change(s)/variant(s). The threshold assignment of OR > 2 used to evaluate polygenic risk during the PRS calculation is discussed in Hao et al. (2022).

An example PRS results table for atrial fibrillation as generated through this protocol is shown in Table 18. Each results table indicates the common disease condition analyzed, the general population lifetime risk of the disease, the simplified polygenic risk result, a summary of the disease risk interpretation, and general disease information as summarized from the literature. The general lifetime risk of disease reflects the risk of developing the particular disease within a certain time interval, which may vary depending on the research study as well the age of a given participant. The simplified polygenic risk result indicates whether the participant is at an increased or average polygenic risk given the calculated and adjusted PRS. The risk interpretation summary provides more details about this polygenic risk result, including the number of loci used for the PRS calculation and the research study used as reference for determining increased and average risk. The general disease information summary provides a brief overview of the common disease being analyzed, the percentage of affected individuals within the general Canadian population, and average lifetime risk, as well as summarizing other non‐genetics related factors that may contribute to the risk of developing the disease. The full list of references included on the PRS tables can be found in File S6 in Supporting Information.

Table 18.

Example of Polygenic Risk Score Results for Atrial Fibrillation

Disease General population lifetime risk Your result
Atrial fibrillation 1 in 4 Average polygenic risk
Risk interpretation: The individual's calculated polygenic risk score, derived from 6730541 loci, has NOT been associated with an increased risk for atrial fibrillation, defined as a greater than 2‐fold risk. Scores that fall among the top 7% of polygenic risk score values are associated with a greater than 2‐fold risk of developing atrial fibrillation among 409,258 participants of British ancestry when compared to the average individual (Khera_2018_30104762).
Disease information: Atrial fibrillation (sometimes called Afib) is a type of irregular heart rhythm (arrhythmia) that affects the upper chambers of the heart (atria) and causes the heart to beat very quickly. This may lead to heart failure and stroke. Atrial fibrillation is the most common type of heart arrhythmia, affecting 1%‐2% of the general population and approximately 200,000 Canadians (Andrade_2020_33191198; Heart and Stroke Foundation of Canada_2020_Atrial fibrillation). Data from an American cohort of men and women (N = 8725) obtained between 1968‐1999 estimated that the lifetime risk (up to 95 years of age) to develop atrial fibrillation was 1 in 4 for men and women 40 years of age or older and 1 in 6 for men and women 40 years of age or older specifically without a previous diagnosis of congestive heart failure or myocardial infarction (Lloyd‐Jones_2004_ 15313941). While genetics plays a role in the risk for atrial fibrillation, it is not the whole picture. There can be many contributors to the development of atrial fibrillation, some of which are coronary artery disease, high blood pressure, diabetes, unhealthy weight, and heavy alcohol consumption.
Open in a new tab

Blood group genotyping is included on the genome report to reveal potential associations between ABO and other blood genotypes and susceptibility to a COVID‐19 infection. For example, studies have shown that individuals with blood group A may be at a higher risk for developing COVID‐19 than those with non‐A blood groups, while individuals with blood group O may actually be more protected against a COVID‐19 infection than those with non‐O blood groups (Severe Covid‐19 GWAS Group et al., 2020). Individuals who are Rhesus‐negative (Rh–) may also have a protective advantage against COVID‐19 versus those who are Rhesus‐positive (Rh+; Ray, Schull, Vermeulen, & Park, 2021). Germline short variants are called from the red blood cell and platelet antigen genes and promoter regions using Genomic Analysis Tool Kit (GATK) v3.7‐0‐gcfedb67 (UnifiedGenotyper and EMIT_ALL_SITES output mode). Sequencing coverage is extracted from the alignment file using BEDTools v2.17.0. Antigen typing using bloodTyper is performed at the relevant allele nt positions corresponding to specific antigens using a 4× calling cutoff and SV detection using a combination of read depth, paired reads, and split read methods. The read depth method evaluates copy number and can detect SV changes such as deletions or replacements. The paired read method in which aligned regions are abnormally spaced is useful for observing breakpoints and evaluating read alignments, deletions, and duplications. The split read method detects when a single read matches alignments from disparate, nonsequential locations, indicating breakpoint transition (Halls et al., 2020, Lane et al., 2018).

Example blood group genotyping results tables are shown in Table 19. The first table summarizes the red blood cell (RBC) antigen predictions, including ABO blood type and RhD status. An extended list of RBC antigens is also predicted, including any rare RBC antigens within a given participant. The second table summarizes human platelet antigen (HPA) predictions, similarly including any extended or rare platelet antigen results for the participant. The third table discussion summarizes the RBC and HPA prediction methodology, including any potential implications of particular antigen predictions that may result in complications during pregnancy or blood transfusions for the participant. Finally, the blood product donation summary indicates potential uncommon or rare antigen blood donation candidacy based on predicted genotypes, as well as potential compatibility for blood donations in participants who are determined to have rare blood type results (Halls et al., 2020, Lane et al., 2018).

Table 19.

Example of Blood Group Genotyping Results

RBC Antigens
ABO (RhD) blood type O (RhD+)
Extended RBC antigen genotyping M+, N‐, Mi(a‐), S‐, s+, U+, He‐, P1+, C+, E+, c+, e+, C(W‐), C(X‐), V‐, VS‐, Lu(a‐), Lu(b+), K‐, k+, Kp(a‐), Kp(b+), Js(a‐), Js(b+), Kp(c‐), Fy(a+), Fy(b+), Jk(a+), Jk(b+), Di(a‐), Di(b+), Wr(a‐), Wr(b+), Sc1+, Sc2‐, Do(a+), Do(b+), Hy+, Jo(a+), Co(a+), Co(b‐), LW(a+), LW(b‐), Cr(a+), Kn(a+), Kn(b‐), McC(a+), Yk(a+), McC(b‐), KCAM+, KDAS‐, Vel+
Rare RBC antigen(s)
Platelet antigens
Extended platelet antigen genotyping HPA‐1a+, HPA‐1b‐, HPA‐2a+, HPA‐2b‐, HPA‐3b+, HPA‐3a‐, HPA‐4a+, HPA‐4b‐, HPA‐5a+, HPA‐5b‐, HPA‐6bw‐, HPA‐15b+, HPA‐15a‐
Rare platelet antigen(s)
DISCUSSION
These red blood cell (RBC) and human platelet antigen (HPA) predictions are based on published genotype to phenotype correlations for the alleles present. Some antigens have also been serologically determined using traditional blood typing methods. During pregnancy or transfusion, alloantibodies to blood group antigens and platelet antigens can form against foreign RBCs that contain immunogenic blood group and platelet antigens that the recipient is missing. These alloantibodies can cause clinically important complications during future transfusions and pregnancy.
Blood product donation
This individual would be a desirable donor for the following uncommon antigen negative change(s) found in <=40% of donors: N‐, HPA‐3a‐, HPA‐15a‐. This individual does not have a rare antigen negative changes.
Open in a new tab

Viral lineage was also determined for each participant and included on the final genome report. Viral consensus sequences are determined using the CanCOGeN‐VirusSeq standardized analysis pipeline based on freebayes v1.7 (https://github.com/freebayes/freebayes, RRID:SCR_010761). Quality control is performed using ncov‐tools (https://github.com/jts/ncov‐tools). The viral lineages reported are predictions and not definitive. They are thus representative of the current knowledge of the lineage at the time of issuing the report to participants.

Critical Parameters

HLA calls made by WGS may not be as accurate as PCR‐based sequencing methods; therefore, clinical HLA testing is recommended to confirm HLA assignments. Phase of HLA genotypes may not be determined through this protocol. Reported HLA‐disease associations are correlative, not causative. Having a specific HLA genotype is not sufficient for developing the associated disease. Further, the in‐house‐developed HLA database does not contain an exhaustive list of all possible disease associations published in the literature. Reported pharmacogenomic associations are risks as determined by current CPIC guidelines. The presence of a specific genotype‐medication metabolizer relationship does not necessarily mean that an adverse reaction to the medication will occur. Reported metabolizer phenotypes and dosing recommendations are meant to be used as guidelines for certified health professionals. Medication prescriptions and dosage should ultimately be left to the discretion of the appropriate healthcare professional/provider. Reported ancestry estimates are generated using datasets containing genomic data from populations of primarily European ancestry; therefore ancestry estimates may not be as accurate for individuals of Non‐White/Non‐European ancestry. The predictive value of a PRS may also not be as accurate for individuals of Non‐White/Non‐European ancestry due to similar dataset limitations. In some cases, reported ancestry may not accurately represent self‐reported ancestry.

Troubleshooting

Possible problems, causes, and suggested solutions are provided in Table 20. In general, troubleshoot possible input errors through log files in output folders.

Table 20.

Troubleshooting

Problem Possible cause Solution
Installation System incompatibilities—setup and script configuration differs between operating systems (e.g., PLINK, samtools, bcftools, htslib) Download the most recent version of each package for the correct OS—Mac and Linux distributions recommended.

Use Docker images for the ease of running each of the pipelines without manual installation and configuration of packages.

OR

Follow the manual instructions step‐by‐step and refer to each of the package distributors for possible installation issues.

Input files Incompatibility between BAM/VCF/FASTQ files Ensure all files are generated from the same FASTQ input, and aligned to the same reference genome.
Truncation or corruption errors Verify the integrity of your FASTQ, VCF, BAM files.
BAMs are not aligned to GRCh37/hg19 BAMs must be aligned to GRCh37 (hg19) for HLA genotyping using HLA‐VBSeq.
VCF files generated outside of GATK Best Practices VCF generation using GATK Best Practices pipeline is recommended.
Chromosome notation (e.g., “1” vs. “chr1”) Ensure chromosome prefixes are compatible between reference fasta and input files.
Computing resources When your local system does not meet the hardware requirements needed to run each of the pipelines and Docker images

Access pipelines through an HPC resource.

OR

If using Docker images, opt to download the Github repository and run script on non‐containerized environment.
Package does not exist Package directory is not included in the PATH variable Follow instructions for given OS system on adding package directory in PATH variable.
Configuration has not been run successfully Look into your system C compiler, and package distributors for successful configuration.
Open in a new tab

Time Considerations

Time and compute considerations can be found in Table 21.

Table 21.

Time and Compute Considerations

Basic protocol Runtime a System resources required
HLA 2 hr N/A
Ancestry 12 hr Minimum 8 GB RAM
Pharmacogenomics 1 hr N/A
HLA (Docker image) N/A Minimum 16 GB RAM, 4 CPUs
Ancestry (Docker image) N/A Minimum 16 GB RAM, 4 CPUs
Pharmacogenomics (Docker) N/A Minimum 16 GB RAM, 4 CPUs
Open in a new tab
a

Runtime is based on a 64 GB RAM M1 MAX system

Author Contributions

Erika Frangione: conceptualization, formal analysis, methodology, software, writing original draft, writing review and editing; Monica Chung: conceptualization, formal analysis, methodology, software, writing original draft, writing review and editing; Selina Casalino: conceptualization, formal analysis, methodology, writing original draft, writing review and editing; Georgia MacDonald: conceptualization, methodology, writing original draft, writing review and editing; Sunakshi Chowdhary: project administration, writing review and editing; Chloe Mighton: project administration, writing review and editing; Hanna Faghfoury: project administration; Yvonne Bombard: project administration; Lisa Strug: data curation; Trevor Pugh: data curation; Jared Simpson: data curation; Limin Hao: methodology, software, writing review and editing; Matthew Lebo: methodology, software, writing review and editing; William J. Lane: formal analysis, methodology, software, writing review and editing; Jennifer Taher: conceptualization, methodology, project administration, supervision, writing review and editing; Jordan Lerner‐Ellis: conceptualization, methodology, project administration, supervision, writing original draft, writing review and editing; Saranya Arnoldo: project administration; Navneet Aujla: formal analysis; Erin Bearss: project administration; Alexandra Binnie: project administration; Bjug Borgundvaag: project administration; Laurent Briollais: project administration; Howard Chertkow: project administration; Marc Clausen: project administration; Marc Dagher: project administration; Luke Devine: project administration; David Di Iorio: Formal analysis; Steven Marc Friedman: project administration; Chun Yiu Jordan Fung: project administration; Anne‐Claude Gingras: project administration; Lee W. Goneau: project administration; Deepanjali Kaushik: project administration; Zeeshan Khan: project administration; Elisa Lapadula: project administration; Tiffany Lu: project administration; Tony Mazzulli: project administration; Allison McGeer: project administration; Shelley L McLeod: project administration; Gregory Morgan: Formal analysis; David Richardson: project administration; Seth Stern: project administration; Ahmed Taher: project administration; Iris Wong: project administration; Natasha Zarei: project administration.

Conflict of Interest

Matthew S. Lebo and Limin Hao are employed by a non‐for‐profit molecular diagnostics lab (Laboratory for Molecular Medicine at Mass General Brigham) that offers genome screening and polygenic risk score assessments. William J. Lane is a member of the Scientific Advisory Board of CareDx, Inc., member of the Next Generation Sequencing Steering Committee of One Lambda, Inc., and his institution is a founding member of the Blood Transfusion Genomics Consortium (BGC) that has received fees from Thermo Fisher Scientific Inc. to help co‐develop a high density DNA genotyping array.

Supporting information

Supplementary File 1: GENCOV_Example_Report.docx. An example of a comprehensive GENCOV genome report to be distributed to participants and physicians.

Click here for additional data file. (301.1KB, docx)

Supplementary File 2: HLA_Database.xlsx. An in‐house developed database of HLA‐disease associations for comparison to identified HLA genotypes across participants.

Click here for additional data file. (35.4KB, xlsx)

Supplementary File 3: Dosage_Recommendations.xlsx. Table of pharmacogenes evaluated through Basic Protocol 3 with formal recommendations compiled from PharmGKB.

Click here for additional data file. (31.6KB, xlsx)

Supplementary File 4: Sample_Appendix.xlsx. Example of a generated pharmacogenomics recommendations appendix outlined in Basic Protocol 3.

Click here for additional data file. (13.7KB, xlsx)

Supplementary File 5: Gene‐Disease_Panel_Database.xlsx. Comprehensive database of medically actionable gene panels used for variant filtration and assessment following ACMG guidelines.cpz1534‐sup‐0001‐FileS1.

Click here for additional data file. (2.3MB, xlsx)

Acknowledgments

This work was supported by funding from the Canadian Institutes of Health Research (Funding Reference Number VR4‐172753, 461170, 461304). We would also like to acknowledge CanCOGen for Host and Viral genome sequencing (https://genomecanada.ca/challenge‐areas/cancogen/).

HostSeq data storage and sharing sub‐Committee

Steven Jones (Chair, BC Cancer Agency)

Inanc Birol (University of British Columbia)

Guillaume Bourque (McGill University)

Lisa Strug (Hospital for Sick Children, Toronto)

Joe Whitney (Hospital for Sick Children)

Bhooma Thiruv (Hospital for Sick Children)

Frangione, E. , Chung, M. , Casalino, S. , MacDonald, G. , Chowdhary, S. , Mighton, C. , Faghfoury, H. , Bombard, Y. , Strug, L. , Pugh, T. , Simpson, J. , Hao, L. , Lebo, M. , Lane, W. J. , Taher, J. , Lerner‐Ellis, J. , & GENCOV Study Workgroup . (2022). Genome reporting for healthy populations—pipeline for genomic screening from the GENCOV COVID‐19 study. Current Protocols, 2, e534. doi: 10.1002/cpz1.534

*GENCOV Study Workgroup: Saranya Arnoldo1,2, Navneet Aujla1, Erin Bearss3, Alexandra Binnie2, Bjug Borgundvaag1,3, Laurent Briollais8, Howard Chertkow4, Marc Clausen5, Marc Dagher1,6, Luke Devine1,3, David Di Iorio1, Steven Marc Friedman3,7, Chun Yiu Jordan Fung3,8, Anne‐Claude Gingras1,3,8, Lee W. Goneau9, Deepanjali Kaushik2, Zeeshan Khan10, Elisa Lapadula3,8, Tiffany Lu3, Tony Mazzulli1,3, Allison McGeer1,3,8, Shelley L. McLeod1,3, Gregory Morgan1,3,8, David Richardson2, Seth Stern10, Ahmed Taher1,7,10, Iris Wong10, and Natasha Zarei10

1University of Toronto, Toronto, Ontario, Canada

2William Osler Health System, Toronto, Ontario, Canada

3Mount Sinai Hospital, Sinai Health, Toronto, Ontario, Canada

4Baycrest Health Sciences, Toronto, Ontario, Canada

5Unity Health Toronto, Toronto, Ontario, Canada

6Women's College Hospital, Toronto, Ontario, Canada

7University Health Network, Toronto, Ontario, Canada

8Lunenfeld‐Tanenbaum Research Institute, Toronto, Ontario, Canada

9Dynacare Medical Laboratories, Toronto, Ontario, Canada

10Mackenzie Health, Richmond Hill, Ontario, Canada

Published in the Human Genetics section

Data Availability Statement

Data openly available in a public repository that issues datasets with DOIs.

Literature Cited

  1. Auton, A. , Brooks, L. D. , Durbin, R. M. , Garrison, E. P. , Kang, H. M. , … Abecasis, G. R. , 1000 Genomes Project Consortium . (2015). A global reference for human genetic variation. Nature, 526(7571), 68–74. doi: 10.1038/nature15393 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alexander, D. H. , Novembre, J. , & Lange, K. (2009). Fast model‐based estimation of ancestry in unrelated individuals. Genome Research, 19(9), 1664–1664. doi: 10.1101/gr.094052.109 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anastassopoulou, C. , Gkizarioti, Z. , Patrinos, G. P. , & Tsakris, A. (2020). Human genetic factors associated with susceptibility to SARS‐CoV‐2 infection and Covid‐19 disease severity. Human Genomics, 14(1), 40. doi: 10.1186/s40246-020-00290-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Crews, K. R. , Gaedigk, A. , Dunnenberger, H. M. , Leeder, J. S. , Klein, T. E. , Caudle, K. E. , … Skaar, T. C. (2014). Clinical pharmacogenetics implementation consortium guidelines for cytochrome p450 2d6 genotype and codeine therapy: 2014 update. Clinical Pharmacology & Therapeutics, 95(4), 376–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. DiStefano, M. T. , Goehringer, S. , Babb, L. , Alkuraya, F. S. , Amberger, J. , Amin, M. , … Rehm, H. L. (2022). The gene curation coalition: A global effort to harmonize gene–disease evidence resources. Genetics in Medicine, 24(8), 1732–1742. doi: 10.1016/j.gim.2022.04.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Eberle, M. A. , Fritzilas, E. , Krusche, P. , Källberg, M. , Moore, B. L. , Bekritsky, M. A. , … Bentley, D. R. (2017). A reference data set of 5.4 million phased human variants validated by genetic inheritance from sequencing a three‐generation 17‐member pedigree. Genome Research, 27(1), 157–164. doi: 10.1101/gr.210500.116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Gao, J. , Zhu, C. , Zhu, Z. , Tang, L. , Liu, L. , Wen, L. , & Sun, L. (2019). The human leukocyte antigen and genetic susceptibility in human diseases. Journal of Bio‐X Research, 2(3), 112–120. doi: 10.1097/JBR.0000000000000044 [DOI] [Google Scholar]
  8. Halls, J. , Vege, S. , Simmons, D. P. , Aeschlimann, J. , Bujiriri, B. , Mah, H. H. , … Lane, W. J. (2020). Overcoming the challenges of interpreting complex and uncommon RH alleles from whole genomes. Vox Sanguinis, 115(8), 790–801. doi: 10.1111/vox.12963 [DOI] [PubMed] [Google Scholar]
  9. Hamosh, A. , Scott, A. F. , Amberger, J. , Valle, D. , & McKusick, V. A. (2000). Online mendelian inheritance in man (OMIM). Human Mutation, 15(1), 57–61. doi: 10.1002/(SICI)1098-1004(200001)15:13.0.CO;2-G [DOI] [PubMed] [Google Scholar]
  10. Hao, L. , Kraft, P. , Berriz, G. F. , Hynes, E. D. , Koch, C. , Kumar, P. K. V. , … Lebo, M. S. (2022). Development of a clinical polygenic risk score assay and reporting workflow. Nature Medicine, 28, 1006–1013. doi: 10.1038/s41591-022-01767-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Holm, I. A. , Agrawal, P. B. , Ceyhan‐Birsoy, O. , Christensen, K. D. , Fayer, S. , Frankel, L. A. , … Beggs, A. H. (2018). The babyseq project: Implementing genomic sequencing in newborns. BMC Pediatrics, 18(1), 225. doi: 10.1186/s12887-018-1200-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Huyghe, J. R. , Bien, S. A. , Harrison, T. A. , Kang, H. M. , Chen, S. , Schmit, S. L. , … Peters, U. (2019). Discovery of common and rare genetic risk variants for colorectal cancer. Nature Genetics, 51(1), 76–87. doi: 10.1038/s41588-018-0286-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Johnson, J. A. , Caudle, K. E. , Gong, L. , Whirl‐Carrillo, M. , Stein, C. M. , Scott, S. A. , … Wadelius, M. (2017). Clinical pharmacogenetics implementation consortium (CPIC) guideline for pharmacogenetics‐guided warfarin dosing: 2017 update. Clinical Pharmacology and Therapeutics, 102(3), 397–404. doi: 10.1002/cpt.668 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Khera, A. V. , Chaffin, M. , Aragam, K. G. , Haas, M. E. , Roselli, C. , Choi, S. H. , … Kathiresan, S. (2018). Genome‐wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nature Genetics, 50(9), 1219–1224. doi: 10.1038/s41588-018-0183-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Landrum, M. J. , Lee, J. M. , Benson, M. , Brown, G. R. , Chao, C. , Chitipiralla, S. , … Maglott, D. R. (2018). Clinvar: Improving access to variant interpretations and supporting evidence. Nucleic Acids Research, 46(D1), D1062–D1067. doi: 10.1093/nar/gkx1153 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Lane, W. J. , Westhoff, C. M. , Gleadall, N. S. , Aguad, M. , Smeland‐Wagman, R. , Vege, S. , … MedSeq Project . (2018). Automated typing of red blood cell and platelet antigens: A whole‐genome sequencing study. The Lancet Haematology, 5(6), e241–e251. doi: 10.1016/S2352-3026(18)30053-X [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Lee, S. B. , Wheeler, M. M. , Thummel, K. E. , & Nickerson, D. A. (2019). Calling star alleles with Stargazer in 28 pharmacogenes with whole genome sequences. Clinical Pharmacology & Therapeutics, 106(6), 1328–1337. doi: 10.1002/cpt.1552 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Li, H. , & Durbin, R. (2009). Fast and accurate short read alignment with Burrows‐Wheeler transform. Bioinformatics, 25(14), 1754–1760. doi: 10.1093/bioinformatics/btp324 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Matzaraki, V. , Kumar, V. , Wijmenga, C. , & Zhernakova, A. (2017). The MHC locus and genetic susceptibility to autoimmune and infectious diseases. Genome Biology, 18(1), 76. doi: 10.1186/s13059-017-1207-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Mavaddat, N. , Pharoah, P. D. P. , Michailidou, K. , Tyrer, J. , Brook, M. N. , Bolla, M. K. , … Garcia‐Closas, M. (2015). Prediction of breast cancer risk based on profiling with common genetic variants. Journal of the National Cancer Institute, 107(5), djv036. doi: 10.1093/jnci/djv036 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Mckenna, A. , Hanna, M. , Banks, E. , Sivachenko, A. , Cibulskis, K. , Kernytsky, A. , … Depristo, M. A. (2010). The genome analysis toolkit: A MapReduce framework for analyzing next‐generation DNA sequencing data. Genome Research, 20(9), 1297–1303. doi: 10.1101/gr.107524.110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Mersha, T. B. , & Abebe, T. (2015). Self‐reported race/ethnicity in the age of genomic research: Its potential impact on understanding health disparities. Human Genomics, 9, 1. doi: 10.1186/s40246-014-0023-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Meyer, H. (2021a). Processing Hapmap III reference data for ancestry estimation. Available at https://meyer‐lab‐cshl.github.io/plinkQC/articles/HapMap.html
  24. Meyer, H. (2021b). Ancestry estimation based on reference samples of known ethnicities. Available at https://meyer‐lab‐cshl.github.io/plinkQC/articles/AncestryCheck.html
  25. Michailidou, K. , Lindström, S. , Dennis, J. , Beesley, J. , Hui, S. , Kar, S. , … Easton, D. F. (2017). Association analysis identifies 65 new breast cancer risk loci. Nature, 551(7678), 92–94. doi: 10.1038/nature24284 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Miller, D. T. , Lee, K. , Chung, W. K. , Gordon, A. S. , Herman, G. E. , Klein, T. E. , … Martin, C. L. (2021). ACMG sf v3.0 list for reporting of secondary findings in clinical exome and genome sequencing: A policy statement of the American College of Medical Genetics and Genomics (ACMG). Genetics in Medicine, 23(8), 1381–1390. doi: 10.1038/s41436-021-01172-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mimori, T. , Yasuda, J. , Kuroki, Y. , Shibata, T. F. , Katsuoka, F. , Saito, S. , … Yamamoto, M. (2019). Construction of full‐length Japanese reference panel of class I HLA genes with single‐molecule, real‐time sequencing. Pharmacogenomics Journal, 19, 136–146. doi: 10.1038/s41397-017-0010-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Purcell, S. , Neale, B. , Todd‐Brown, K. , Thomas, L. , Ferreira, M. A. R. , Bender, D. , … Sham, P. C. (2007). PLINK: A toolset for whole‐genome association and population‐based linkage analysis. American Journal of Human Genetics, 81, 559–575. doi: 10.1086/519795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Ray, J. G. , Schull, M. J. , Vermeulen, M. J. , & Park, A. L. (2021). Association between ABO and RH blood groups and SARS‐CoV‐2 infection or severe Covid‐19 illness. Annals of Internal Medicine, 174(3), 308–315. doi: 10.7326/M20-4511 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Reble, E. , Salazar, M. G. , Zakoor, K.‐R. , Khalouei, S. , Clausen, M. , Kodida, R. , … Bombard, Y. (2021). Beyond medically actionable results: An analytical pipeline for decreasing the burden of returning all clinically significant secondary findings. Human Genetics, 140(3), 493–504. doi: 10.1007/s00439-020-02220-9 [DOI] [PubMed] [Google Scholar]
  31. Rehm, H. L. , Berg, J. S. , Brooks, L. D. , Bustamante, C. D. , Evans, J. P. , Landrum, M. J. , … Watson, M. S. (2015). Clingen—the clinical genome resource. New England Journal of Medicine, 372(23), 2235–2242. doi: 10.1056/NEJMsr1406261 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Richards, S. , Aziz, N. , Bale, S. , Bick, D. , Das, S. , Gastier‐Foster, J. , … ACMG Laboratory Quality Assurance Committee . (2015). Standards and guidelines for the interpretation of sequence variants: A joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genetics in Medicine, 17(5), 405–424. doi: 10.1038/gim.2015.30 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Schumacher, F. R. , Olama, A. A. A. , Berndt, S. I. , Benlloch, S. , Ahmed, M. , Saunders, E. J. , … Eeles, R. A. (2018). Association analyses of more than 140,000 men identify 63 new prostate cancer susceptibility loci. Nature Genetics, 50(7), 928–936. doi: 10.1038/s41588-018-0142-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Ellinghaus, D. , Degenhardt, F. , Bujanda, L. , Buti, M. , Albillos, A. , … Karlsen, T. H. , Severe Covid‐19 GWAS Group . (2020). Genomewide association study of severe Covid‐19 with respiratory failure. The New England Journal of Medicine, 383(16), 1522–1534. doi: 10.1056/NEJMoa2020283 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Solomon, B. D. , Nguyen, A.‐D. , Bear, K. A. , & Wolfsberg, T. G. (2013). Clinical genomic database. Proceedings National Academy of Science USA, 110(24), 9851–9855. doi: 10.1073/pnas.1302575110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sze, S. , Pan, D. , Nevill, C. R. , Gray, L. J. , Martin, C. A. , Nazareth, J. , … Pareek, M. (2020). Ethnicity and clinical outcomes in Covid‐19: A systematic review and meta‐analysis. EClinicalMedicine, 29, 100630. doi: 10.1016/j.eclinm.2020.100630 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Taher, J. , Mighton, C. , Chowdhary, S. , Casalino, S. , Frangione, E. , Arnoldo, S. , … Lerner‐Ellis, J. (2021). Implementation of serological and molecular tools to inform Covid‐19 patient management: Protocol for the GENECOV prospective cohort study. BMJ Open, 11(9), e052842. doi: 10.1136/BMJOPEN-2021-052842 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Altshuler, D. M. , Gibbs, R. A. , Peltonen, L. , Altshuler, D. M. , Gibbs, R. A. , … McEwen, J. E. , The International HapMap 3 Consortium . (2010). Integrating common and rare genetic variation in diverse human populations. Nature, 467(7311), 52–58. doi: 10.1038/nature09298 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Thorn, C. F. , Klein, T. E. , & Altman, R. B. (2013). PharmGKB: The pharmacogenomics knowledge base. Methods in Molecular Biology, 1015, 311–320. doi: 10.1007/978-1-62703-435-7_20 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Whirl‐Carrillo, M. , Huddart, R. , Gong, L. , Sangkuhl, K. , Thorn, C. F. , Whaley, R. , & Klein, T. E. (2021). An evidence‐based framework for evaluating pharmacogenomics knowledge for personalized medicine. Clinical Pharmacology & Therapeutics, 110(3), 563–572. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Zeberg, H. , & Pääbo, S. (2020). The major genetic risk factor for severe Covid‐19 is inherited from Neanderthals. Nature, 587(7835), 610–612. doi: 10.1038/s41586-020-2818-3 [DOI] [PubMed] [Google Scholar]

Internet Resources

COVID‐NET: Covid‐19‐associated hospitalization surveillance network, centers for disease control and prevention (2021).

  1. https://gis.cdc.gov/grasp/covidnet/covid19_3.html

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary File 1: GENCOV_Example_Report.docx. An example of a comprehensive GENCOV genome report to be distributed to participants and physicians.

Click here for additional data file. (301.1KB, docx)

Supplementary File 2: HLA_Database.xlsx. An in‐house developed database of HLA‐disease associations for comparison to identified HLA genotypes across participants.

Click here for additional data file. (35.4KB, xlsx)

Supplementary File 3: Dosage_Recommendations.xlsx. Table of pharmacogenes evaluated through Basic Protocol 3 with formal recommendations compiled from PharmGKB.

Click here for additional data file. (31.6KB, xlsx)

Supplementary File 4: Sample_Appendix.xlsx. Example of a generated pharmacogenomics recommendations appendix outlined in Basic Protocol 3.

Click here for additional data file. (13.7KB, xlsx)

Supplementary File 5: Gene‐Disease_Panel_Database.xlsx. Comprehensive database of medically actionable gene panels used for variant filtration and assessment following ACMG guidelines.cpz1534‐sup‐0001‐FileS1.

Click here for additional data file. (2.3MB, xlsx)

Data Availability Statement

Data openly available in a public repository that issues datasets with DOIs.

Articles from Current Protocols are provided here courtesy of Wiley

RESOURCES