New Resources to Identify Characterized DNA Reference Materials for PGx and HLA Testing: The Genetic Testing Reference Material (GeT-RM) Program PGx Search Tool and GeT-RM Consolidated PGx and HLA Table

Abstract

Regulations, accreditation standards, and professional guidance require laboratories to use reference materials for assay development, validation, quality control, and proficiency testing of clinical genetic tests. There are, however, few publicly available reference materials for most genetic tests. To address this issue, the Centers for Disease Control and Prevention’s Genetic Testing Reference Material Program (GeT-RM), the Coriell Institute for Medical Research, and the genetic testing community have conducted 19 studies, including nine for pharmacogenetic (PGx) and Human Leukocyte Antigen (HLA) testing, to create characterized, renewable, and publicly available DNA samples for use as reference materials. Because new PGx alleles are frequently identified, and allele designations change over time, many samples were reanalyzed for the same gene(s) in subsequent GeT-RM studies. These studies utilized more comprehensive and sensitive methods and panels that examined additional single nucleotide variants (SNVs) and/or star alleles to expand and update the consensus genotypes. Up to date information is available in two newly established resources: the GeT-RM Consolidated PGx and HLA Table and the GeT-RM PGx Search Tool. These resources contain all available PGx and HLA genotypes for 363 publicly available samples characterized during nine GeT-RM PGx or HLA studies for 34 genes/loci in a consolidated and searchable format.

Introduction

Reference materials are critical for many activities in clinical laboratories, including quality control, development and validation of tests, proficiency testing, validating variant calling algorithms, and inter-laboratory standardization, and are needed to comply with regulations, accreditation standards, and professional guidance^1–5 (American College of Medical Genetics and Genomics https://www.acmg.net/PDFLibrary/ACMG%20Technical%20Lab%20Standards%20Section%20G.pdf, last accessed 7/12/2024, Washington State Legislature, http://app.leg.wa.gov/WAC/default.aspx?cite=246-338-090, last accessed 7/12/2024, College of American Pathologists (Northfield, IL) https://www.cap.org/ last accessed 7/12/2024, New York State Clinical Laboratory Evaluation Program, https://www.wadsworth.org/regulatory/clep, last accessed 7/12/2024, MMWR https://www.cdc.gov/mmwr/preview/mmwrhtml/rr5806a1.htm, last accessed 7/12/2024).

Clinical laboratories often develop pharmacogenetic and other genetic tests as laboratory developed tests or procedures (LDT or LDP), and use residual patient samples, genomic DNA from cell lines, or synthetic DNA containing variants of interest as reference materials. However, despite the regulatory and professional guidelines requiring their use, there are few, if any, publicly available reference materials for most clinical genetic tests.

To address this issue, the CDC created the Genetic Testing Reference Materials Program (GeT-RM) in 2004 (GeT-RM, https://www.cdc.gov/lab-quality/php/get-rm/index.html, last accessed 11/4/2024).⁶ The GeT-RM is a collaborative program whose goal is to improve the availability of genomic DNA reference materials for genetic testing. GeT-RM works with a variety of partners, including clinical and research laboratories, in vitro diagnostics (IVD) manufacturers, professional organizations, patient advocacy groups, and the Coriell Institute for Medical Research to identify reference material needs and create publicly available and renewable genomic DNA reference materials.

GeT-RM has created cell-line based genomic DNA reference materials for a large number of clinically important genes for hereditary genetic disorders, including cystic fibrosis, fragile X, Rett syndrome, and Duchenne muscular dystrophy^7–10, 11 Human Leukocyte Antigen (HLA) loci¹¹, and many pharmacogenes and loci.^12–19 Each sample is experimentally characterized in two or more laboratories using a variety of analytical methods, including targeted genotyping and Sanger sequencing. For some genes, experimental testing also includes copy number variants (CNVs) and structural variation (SVs) analysis or phasing of variants into star alleles.²⁰ In addition, publicly available high-coverage (30x) whole genome sequencing (WGS) data from the 1000 Genomes Project (1kGP)²¹ or 10x Genomics Linked-Read data (Illumina https://github.com/Illumina/Polaris/wiki/HiSeqX-PGx-Cohort last accessed 8/19/2024) was also analyzed. Results were assessed for quality, discordances, and determination of a consensus genotype for each sample. All characterized genomic DNA samples are publicly available from the National Institute of General Medical Sciences (NIGMS) Human Genetic Cell Repository or the National Human Genome Research Institute (NHGRI) Sample Repository for Human Genetic Research, which are both housed at the Coriell Institute for Medical Research.

The GeT-RM has conducted nine studies over the last 14 years to create reference materials for pharmacogenetic and HLA testing.^11–19 In many cases, samples were reanalyzed for the same gene(s) in subsequent studies using more comprehensive and sensitive methods and panels with additional single nucleotide variants (SNVs) and/or star alleles that were not included in previous characterizations along with other data that became publicly available. This approach allowed updating and expanding consensus genotypes for a considerable number of samples. Often, additional samples were added in later characterization studies to identify SNVs and star alleles that were defined after the original study. In addition, samples characterized during the earliest studies may also have been assigned outdated designations as allele definitions evolve over time and new alleles are identified frequently. These allele designations were updated during subsequent studies.

We created two resources, the GeT-RM Consolidated PGx and HLA Table as well as a searchable database tool, GeT-RM PGx Search, to provide the most up to date pharmacogenetic genotypes for each characterized sample. Both resources provide information about 363 DNA samples that were characterized during nine GeT-RM PGx or HLA studies for 34 genes/loci including CYP2C9, CYP2C19, CYP3A4, CYP3A5, CYP2D6, TPMT, NUDT15, DPYD, and 11 HLA loci.^11–19

Methods

GeT-RM samples

For each GeT-RM study, genomic DNA from cell lines identified by study team members to contain variants of interest were selected from the NIGMS Human Genetic Cell Repository and/or the NHGRI Sample Repository for Human Genetic Research at the Coriell Institute for Medical Research. Each sample was characterized using a variety of methods and test platforms by two or more laboratories.^11–19 Results were assessed for quality, discordances, and determination of consensus genotype for each sample. These data were subsequently used to create two consolidated information resources.

Creation of the GeT-RM Consolidated PGx and HLA Table

Consensus genotypes for all samples (n= 363) characterized during the nine GeT-RM PGx or HLA reference material studies were consolidated into a single, publicly available spreadsheet (in Excel format), referred to as the “GeT-RM Consolidated PGx and HLA” Table (https://www.cdc.gov/lab-quality/php/get-rm/reference-materials.html last accessed 11/4/2024). This consolidated table (Figure 1), includes information regarding the study or studies in which each sample was characterized, the NIGMS or NHGRI Repository sample ID, and each sample’s genotype for the interrogated genes and loci (Table 1). The GeT-RM Consolidated PGx and HLA Table also provides information regarding the availability of Binary Alignment Map (BAM) files, FASTQ files, as well as high-quality 30x coverage WGS data from the 1000 Genomes Project.²¹ Each allele annotation is accompanied by the associated reference(s) that include a detailed description of the methodologies employed to characterize the sample and construct the consensus genotype or star allele annotation. In cases where a given annotation was described in more than one corresponding publication, and the consensus annotation was consistent, all references are included. In cases where a given annotation was described in more than one corresponding publication, and the consensus annotations were inconsistent, the most recent publication and annotation are shown. The GeT-RM consolidated PGx and HLA Table also contains tabs with information about the columns and definitions of terms used, as well as GeT-RM references.

Figure 1.

Screenshot of GeT-RM Consolidated PGx and HLA Table.

Consensus genotypes for 34 gene/loci determined during nine GeT-RM studies are available. References for each genotype are provided in the adjacent column. BAM and FASTQ files are available for some samples, and many have sequence data from the 1000 Genomes Project. The Excel file contains several tabs, shown at the bottom of the figure. The references in the above image do not reflect the citations in this manuscript.

Table 1.

Genes and loci in the GeT-RM Consolidated PGx and HLA Table and GeT-RM PGx Search tools

Gene or Loci	GeT-RM Study(s) PMID
CYP1A1	2662110117
CYP1A2	2662110117
CYP2A6	2662110117
CYP2B6	2662110117
CYP2C8	2662110117, 3513454214
CYP2C9	2088955518, 2662110117, 3513454214, 3402004115
CYP2C19	2088955518, 2662110117, 3513454214, 3402004115
CYP2D6	2088955518, 2662110117, 3140112416
CYP2E1	2662110117
CYP3A4	2662110117, 3735499312
CYP3A5	2662110117, 3735499312
CYP4F2	2662110117
DPYD	2662110117, 3903282219
GSTM1	2662110117
GSTP1	2662110117
GSTT1	2662110117
NAT1	2662110117
NAT2	2662110117
NUDT15	3593134213
SLC15A2	2662110117
SLCO2B1	2662110117
TPMT	2662110117, 3593134213
UGT1A1	2088955518, 2662110117
UGT2B7	2662110117
UGT2B15	2662110117
UGT2B17	2662110117
HLA-A	2995902511
HLA-B	2995902511
HLA-C	2995902511
HLA-DRB1	2995902511
HLA-DRB3	2995902511
HLA-DRB4	2995902511
HLA-DRB5	2995902511
HLA-DQA1	2995902511
HLA-DQB1	2995902511
HLA-DPA1	2995902511
HLA-DPB1	2995902511
CYP2C Cluster NC_000010.10^: g.96405502G>A, rs12777823#	3402004115
GGCX NM_000821.6:c.214+597G>A, rs12714145	3402004115
GGCX NM_000821.6:c.2084+45G>C rs11676382	3402004115
VKORC1 NM_024006.5:c.−1639G>A, rs9923231	2088955518, 2662110117
VKORC1 NM_024006.6:c.106G>A, rs61742245	3402004115
VKORC1 NM_024006.6:c.196G>A, rs72547529	3402004115

^{^}RefSeq https://www.ncbi.nlm.nih.gov/refseq/, last accessed 5/17/2024

^#dbSNP (https://www.ncbi.nlm.nih.gov/snp, last accessed 2/2/2024)

Creation of a searchable database: GeT-RM PGx Search Tool

The GeT-RM Consolidated PGx and HLA Table was used to create a searchable, web-based tool, GeT-RM PGx Search. For each DNA sample included in the GeT-RM Consolidated PGx and HLA Table, the annotation for each characterized pharmacogene was formatted into a single genotype, apart from DPYD for which all variants identified are listed. Most included pharmacogenes are annotated with star (*) alleles. In these cases, the two star alleles are separated by a slash delimiter (“/”); however, a subset of pharmacogenes and regions (e.g., CYP2C Cluster, DPYD, GGCX, SLCO2B1, VKORC1) include variant specific annotations. Finally, HLA gene annotations are included as previously described.¹¹ Reference(s) for each allele annotation are provided as described above.

Results

The GeT-RM Consolidated PGx and HLA Table (https://www.cdc.gov/lab-quality/php/get-rm/reference-materials.html, last accessed 11/4/2024) shows consensus genotypes determined during the nine GeT-RM PGx studies for all 363 samples in an Excel format. This table was used to develop an interactive, web-based search tool, “GeT-RM PGx Search”, which is available on the Coriell Institute for Medical Research website (https://www.coriell.org/GetRM/PGxSearch, last accessed 8/13/2024). Each gene is given a unique key in a Structured Query Language (SQL) Server relational database that is linked to sample level data. Users can query GeT-RM PGx Search using a web-based interface and select a gene of interest from the dropdown menu (Figure 2). GeT-RM PGx Search returns the following: links to PharmVar annotations (https://www.pharmvar.org/, last accessed 7/22/2024), and the NCBI gene entry (if available) (NCBI https://www.ncbi.nlm.nih.gov/gene/?term=, last accessed 8/13/2024), additional documentation regarding allele information and allele definitions, an option to export annotations for the chosen gene to an Excel spreadsheet, and an interactive data display (Figure 3). The number of samples viewable in the interactive table can be adjusted to the user’s preference, and the annotated samples can be sorted by sample ID, description, gene, genotype, reference, product, source, or genetic sex. Each sample ID is hyperlinked to a sample-specific page that includes the detailed overview, characterizations, associated data and publications, and related external links. In addition, the viewable samples can be filtered by star allele, variant(s), or genotype.

Figure 2.

Screenshot of user dropdown menu of genes and loci with available GeT-RM PGx and HLA characterizations.

Figure 3.

Screenshot of example interactive GeT-RM PGx search data display: search for the CYP2D6*17 allele.

The display includes annotation definition information and external links to PharmVar and NCBI. The user can choose the number of viewable samples, sort samples by sample ID, description, gene, genotype, reference, product, source, or genetic sex, export the information to an Excel spreadsheet, and click on any sample hyperlink to access a sample-specific page that includes the detailed overview, characterizations, associated data and publications, and related external links.

Discussion

Reference material organizations including the National Institute of Standards and Technology (NIST, https://www.nist.gov/, last accessed 1/3/2025), the National Institute for Biological Standards and Control (NIBSC, Hertfordshire UK) and the Joint Research Centre (Geel, Belgium) as well as commercial vendors produce reference materials and standards for product control and forensic testing. Availability of reference materials encompassing a range of human genetic variation, however, has been an ongoing challenge for clinical laboratorians. This limited availability has negatively impacted the ability to develop and validate new tests and provide quality control and ongoing quality assurance for existing tests. To address this gap, the GeT-RM has characterized hundreds of publicly available and renewable genomic DNA reference materials, many of these for pharmacogenetic testing. Unlike other PGx annotation approaches^22–24, the GeT-RM requires that each sample be experimentally characterized in two or more laboratories using a variety of analytical methods. These reference materials can then be used by laboratories as they implement clinical practice guidelines such as those developed by the Association for Molecular Pathology PGx Working Group.^25–30

Most pharmacogenes are highly polymorphic^{31, 32} with new alleles and haplotypes still being discovered, especially in non-European poulations;^{33, 34} thus the catalog of haplotype-based (i.e., star allele) definitions continues to grow and is anticipated to be updated regularly.^{20, 35, 36} In addition, there has been a rapid evolution of the technologies and methods available to perform pharmacogenetic testing. Over the last 15 years, pharmacogenetic tests, which started as targeted genotyping and Sanger sequencing assays, have come to incorporate more comprehensive analyses³⁷ including testing for copy number variation, as well as short and long-read NGS.^{38, 39} This has enhanced our understanding of many pharmacogenetic loci, such as CYP2D6, which often requires extensive analysis to resolve its many structural variants.⁴⁰ Historically, simpler assays were only able to detect full CYP2D6 gene deletions, duplications and multiplications, while recent more advanced assays can also detect hybrid genes and more comprehensively characterize complex genotypes. These higher resolution analyses have resulted in the identification and naming of many new CYP2D6 star alleles by PharmVar. While copy number analysis is now commonly performed for CYP2D6, testing for copy number variation is not widely used for other loci. For example, the most recent GeT-RM study utilized copy number analysis and identified a relatively common intragenic deletion in DPYD.¹⁹ Finally, at present, the phase of most pharmacogene variants is determined empirically. As single molecule long read NGS becomes more widely available, it is anticipated that previously analyzed samples may be re-tested to determine phase, which may require further updates to the “gold standard” genotype data in GeT-RM for some samples. Although the reference material DNA does not change over time, the ability to precisely characterize genotype continues to improve.

Since 2010, GeT-RM has performed eight studies characterizing genomic DNA samples for pharmacogenetic loci^12–19 and one study for 11 HLA loci.¹¹ Due to the changing nature of PGx testing and the identification of important new alleles, reference materials for many of the commonly tested PGx genes, including CYP2C9, CYP2C19, CYP2D6, and DPYD, have been characterized in more than one GeT-RM study.^12–19 Some samples were characterized for the same gene during multiple studies, and their consensus genotype was revised based on the information gleaned from analyses using more comprehensive assays. This has resulted in some sample genotypes seeming to change between GeT-RM studies, which may have been interpreted as being “discrepant” and caused confusion amongst users of the reference material samples.

Assignment of PGx genotypes during GeT-RM studies is dependent on the assays used for sample characterization as well as the PharmVar star allele definitions existing at the time of the study. Newly available assay technologies can be used to recharacterize samples to overcome limitations of assays used during earlier studies. For example, genotyping assays, Sanger or short read NGS assays employed in previous studies could often not resolve the phase of variants. This limitation may now be overcome by single molecule long read NGS technologies. In addition, PharmVar may have revised allele definitions and added new star alleles since samples were first characterized. GeT-RM aims to not only update reported genotypes of previously characterized samples by testing with new technologies, but also identify samples with newly discovered star alleles. It is important to note that genotypes listed in the GeT-RM Consolidated PGx and HLA Table are based on the nomenclature available at the time of analysis and may not be consistent with current nomenclature. This explains, for example, why some genotypes contain “letter” extensions (e.g., for CYP1A2). Furthermore, NAT2 was only transferred to PharmVar in 2024 and its nomenclature underwent substantial changes. Future recharacterization of sample materials for NAT2 will not only include testing for many additional star alleles with the latest technologies but will also report genotypes using PharmVar’s current star allele definitions.

There are several bioinformatic tools that can be used to call PGx variants and star alleles for the 1000 Genomes Project samples^{22–24, 41, 42}. These tools can be useful to identify publicly available samples that may contain PGx variants of interest, however the variants in these samples should be confirmed using orthogonal methods prior to use as reference materials.

The GeT-RM Consolidated PGx and HLA Table and the GeT-RM PGx Search tool were created to provide an easily accessible and searchable resource with up-to-date information about publicly available, well-characterized reference DNA samples that can be used to support quality assurance programs of laboratories performing clinical PGx and HLA testing. The GeT-RM Consolidated PGx and HLA Table and GeT-RM PGx Search tool will be updated and synchronized as additional GeT-RM PGx studies are completed. All reference materials characterized by GeT-RM are publicly available from the NIGMS and NHGRI Repositories at the Coriell Institute for Medical Research.