Abstract
Purpose:
Research is underway worldwide to investigate the feasibility, acceptability, and utility of sequencing-based newborn screening. Different methods have been used to select gene-condition pairs for screening, leading to highly inconsistent gene lists across studies.
Methods:
Early Check developed and utilized actionability-based frameworks for evaluating gene-condition pairs for inclusion in newborn panels (Panel 1 - high actionability, Panel 2 - possible actionability). A previously developed framework, the Age-based Semi Quantitative Metric (ASQM), was adapted. Increasing ASQM scores, with a maximum of 15, suggest greater actionability. Wilcoxon tests were performed to compare Panel 1 gene-condition pairs on the Recommended Uniform Screening Panel (RUSP) to non-RUSP pairs.
Results:
In our first round of assessment, Early Check identified 178 gene-condition pairs for inclusion in Panel 1 and 29 for Panel 2. Median ASQM scores of RUSP conditions on Panel 1 was 12 (range 4 to 15) and non-RUSP was 13 (range 9 to 15). Median scores for Panel 2 was 10 (range 6 to 14).
Conclusion:
The Early Check frameworks provide a transparent, semiquantitative, and reproducible methodology for selecting gene-condition pairs for NBS sequencing pilot studies that may inform future integration of genomic sequencing into population-level NBS. Collaborative efforts among newborn sequencing studies to establish shared criteria is needed to enhance cross-study comparisons.
Keywords: newborn screening, newborn sequencing, genome sequencing, clinical actionability, Early Check
INTRODUCTION
The overarching objective of newborn screening (NBS) is to identify children with specific medical conditions in the first days of life who would benefit from interventions to prevent or mitigate morbidity and mortality. Two significant advancements present unprecedented opportunities for the NBS program: the rapid development of new therapies for rare genetic conditions and the enhanced feasibility of using genomic sequencing techniques in large-scale public health applications (1). Significant strides have been made in the treatment of rare monogenic conditions, marked by groundbreaking advances in therapies and genomic technologies that enable early detection, paving the way for improved outcomes for individuals born with these conditions. Members of the NHGRI- and NICHD-funded NSIGHT consortium conducted several exploratory projects to investigate the potential for genomic sequencing in newborns (2). However, ongoing research is essential to address unanswered questions regarding the use of genomic sequencing in NBS, while also addressing challenges related to workforce adequacy and clinical implementation to realize the potential of these technologies (3, 4). Thus, genomic sequencing as a primary method of screening remains on the horizon, with critical unanswered questions about sequencing-based newborn screening.
Currently, about a dozen research pilot studies are enrolling or preparing to enroll participants to investigate the feasibility, acceptability, and utility of genomic sequencing to screen for rare genetic conditions (5). Although these programs have a common goal of informing whether, when, and how to implement population-based newborn sequencing, the recruitment, consent, and return-of-results processes vary considerably. Importantly, methods used to select (at least the initial set of) genes for inclusion on newborn panels also varied, resulting in different research programs screening for somewhat overlapping but largely inconsistent groups of genetic conditions (6). Here, we describe the development of standardized frameworks for selecting gene-condition pairs for inclusion in two newborn sequencing panels for Early Check, a voluntary NBS study available to babies who receive NBS in North Carolina.
Early Check began offering supplemental NBS under a consented model in October 2018 (7–10). The project is led by researchers at RTI International in partnership with the University of North Carolina at Chapel Hill (UNC), and the North Carolina State Laboratory of Public Health (NCSLPH). Birthing parents who deliver a baby in North Carolina and receive standard NBS through the NCSLPH are recruited via multi-modal efforts, including in-person recruitment at UNC and a letter addressed from NCSLPH notifying them of the voluntary Early Check program (11). Interested parents visit an electronic portal (mirrored content available at https://stageportal.earlycheck.org), available in both English and Spanish, to learn more about the study and enroll their babies. Newborns are eligible for enrollment up to 30 days after birth. Screening is performed on the residual state-collected dried blood spots (DBS) after consent is received.
Building on its successful infrastructure, Early Check began planning in 2021 for a significant expansion using genome sequencing (GS) to screen for hundreds of rare monogenic conditions. In addition, a genetic risk score will be implemented to calculate lifetime risk for type 1 diabetes (12). A rigorous and standardized process of expert review and decision-making about the genes and conditions for which to screen preceded enrollment for this phase of Early Check, which opened in September 2023. This publication describing the Early Check gene-condition selection process, frameworks, and gene lists can aid other newborn sequencing studies in selecting gene-condition pairs for screening and may provide a potential model for public health implementation. The future of NBS likely will include genomic sequencing and the selection of which genes to screen will fundamentally impact the potential benefits and risks to infants, families, and society.
METHODS
As part of our planning process to initiate sequencing-based screening, Early Check developed two initial monogenic sequencing panels, informatically selected off a GS backbone. This paper reports on our initial gene panels. Panels continue to grow over time using the methods described here. Panel 1 includes gene-condition pairs with high actionability and early age of action (age at intervention initiation), which makes them comparable to core conditions on the Recommended Uniform Screening Panel (RUSP), a set of conditions the Secretary of the U.S. Department of Health and Human Services (HHS) recommends for inclusion in all state NBS programs (13). All currently recommended RUSP conditions for which sequencing is informative were automatically included on Early Check Panel 1, and other non-RUSP conditions were evaluated for the panel. Panel 2 offers additional screening for gene-condition pairs that do not meet the high thresholds for Panel 1 inclusion but have emerging potential for high actionability. This includes, for example, conditions for which there are clinical trials but not yet approved therapies. All enrolled babies are screened with Panel 1; parents can decide to accept or decline Panel 2 and/or lifetime genetic risk for type 1 diabetes.
The Early Check Gene Panel Working Group (GPWG) was assembled in 2021 with membership consisting of clinician/researchers from RTI and UNC. GPWG members include four clinical geneticists, one pediatrician practicing in primary care, five genetic counselors, one PhD social scientist/genetic counselor, one PhD genetic epidemiologist, one PhD translational genomics researcher, and one clinical psychologist (Panel 2 meetings only). GPWG members have extensive experience in NBS and genomic sequencing. Members include a former Chair of the Advisory Committee for Heritable Disorders in Newborns and Children, the committee that provides advice and recommendations to the Secretary of HHS regarding which conditions should be on the RUSP. Members also include leaders of the National Institutes of Health (NIH)–funded NC NEXUS newborn sequencing project (14) and the Clinical Genome Resource (ClinGen) Actionability Working Group (15). The goal of the GPWG was to develop frameworks for evaluating gene-condition pairs for inclusion in the two monogenic panels and then to use the frameworks to populate the panels.
Early Check Panel 1 Framework
A previously developed framework for evaluating gene-condition pairs for newborn sequencing, the Age-based Semi Quantitative Metric (ASQM) (16) was the starting point for development of the Panel 1 framework for Early Check. The ASQM assesses five criteria of clinical actionability on a scale of 0–3: severity and likelihood of the manifested condition, efficacy and acceptability of the intervention, and knowledge about the gene-condition association and intervention. Increasing ASQM scores, with a maximum of 15, suggest greater actionability. Prior assessment of 34 gene-condition pairs, representing core RUSP conditions at the time, resulted in scores ranging from 4 to 15, with 85% scoring ≥ 12 (16). With that in mind, the GPWG prioritized the consideration of gene-condition pairs with total scores of 13 or higher for evaluation for Panel 1. However, those with scores of ≥ 9 could be considered under specific circumstances: (1) Groups of related conditions were reviewed together (e.g., urea cycle disorders); therefore, conditions with ASQM scores ≤ 12 were reviewed if they belonged to a group with a least one gene-condition pair scoring ≥ 13, and (2) gene-condition pairs could be nominated for review. Nominations are solicited during presentations and at meetings and are accepted from any source. Most were submitted by Early Check investigators or industry sponsors. Nominated gene-condition pairs were typically a focus of the investigator’s personal research or part of industry drug development pipelines. A publicly available nomination form was developed to accept external nominations (https://forms.office.com/r/EJShT5CVkB) from all sources. All but one member of the GPWG was blinded to the source of nominations at the time of discussion and decision-making.
Genes that were associated with core RUSP conditions (e.g., CFTR—cystic fibrosis) were automatically included in Panel 1 regardless of the ASQM score (Figure 1). Genes associated with core RUSP conditions were compiled by a genetic counselor and confirmed by two clinical geneticists (17). For core RUSP conditions with multiple causal genes (severe combined immunodeficiency, congenital hearing loss, and congenital hypothyroidism), the GPWG discussed associated genes and included those that accounted for a significant portion of the condition (e.g., genes that accounted for greater than 2% of SCID) and had definitive or strong gene-disease validity as assessed by ClinGen (18).
Figure 1.
Early Check gene panel frameworks
The published framework categorized the age of action for gene-condition pairs as neonatal, infant, childhood, adolescent, adulthood, or variable (16). A more-specific age of action was used for the Early Check Panel 1 framework. The age of action for inclusion in Panel 1 was set at ≤ 2 years to ensure identification of babies with conditions actionable early in life (Figure 1). The focus on conditions with early actionability was thought to enhance their relevance to parents in the newborn period and enrich for measurable outcomes within a feasible timeframe for research. The “action” had to include surveillance or therapeutic intervention that directly improved health outcomes. Benefits such as family planning or preventing a potential diagnostic odyssey were not determined to be sufficient to include in Panel 1. For gene-condition pairs with a highly variable age of action but well-established genotype-phenotype correlations, there was consensus that specific early-onset variants could be considered individually (19).
Evidence about each gene-condition pair was compiled and curated by a subset of GPWG members. The curated evidence summaries documented the prevalence, penetrance, clinical presentation, age of onset, age of action, follow-up testing needed, and management (supplemental file 1). A simplified version of the literature search developed by the ClinGen Actionability Working Group was used to identify management guidelines from which the recommended age of action was extracted. If management guidelines were not available, the primary literature was used to determine the ages at which surveillance/intervention had been reported to occur. The curated evidence summaries were discussed individually at biweekly GPWG meetings via Zoom followed by an anonymous vote for each gene-condition pair using the Zoom poll feature to determine whether it should be included in Panel 1. At least six voting members, including at least two clinical geneticists, had to be present at all meetings and partake in voting. Gene-condition pairs that received a majority of positive votes were approved for inclusion; unanimous agreement was not required. Panel 1 genes required a gene-disease validity classification of definitive or strong if they had been assessed by ClinGen, although not all genes had been assessed (18). Additional factors such as the technical feasibility of detecting pathogenic variants, the cost and availability of the intervention, and cost-effectiveness of screening were not assessed by this GPWG.
Early Check Panel 2 Framework
To develop a framework for Panel 2, the GPWG selected 18 exemplar conditions from the Newborn Screening Translational Research Network (NBSTRN) RUSP candidate list as potential candidates for inclusion in Panel 2 (20). The GPWG discussed and documented reasons why each would or would not be a good candidate for Panel 2. These discussions formed the foundation of the Panel 2 framework (Figure 1).
The primary inclusion criteria for Panel 2 were ultimately decided to be: 1) conditions with approved treatments that needed additional evidence of the intervention in a presymptomatic newborn to support future inclusion in Panel 1 (e.g., gene therapy for DMD), or 2) conditions without an approved treatment but with interventional clinical trial activity (e.g., investigational gene therapy for Canavan disease). Although ASQM scores were calculated for Panel 2 gene-condition pairs, no minimum ASQM score was set for inclusion in Panel 2. Qualifying clinical trials could be anywhere in the United States, enroll children 2 years of age or younger, and were required to potentially provide clinical benefits to participants, as subjectively assessed by the GPWG. Two sources were used to identify clinical trials: ClinicalTrials.gov and the American Society of Gene + Cell Therapy (ASGCT) Clinical Trials Finder. The GPWG aimed to ensure that parents of children identified with a Panel 2 condition had options, whether through clinical interventions or the opportunity to participate in clinical trials. It was unfeasible to continuously monitor individual trials and adjust gene-condition pairs in Panel 2 based on trials’ initiation and completion. Instead, the existing level of clinical trial activity was considered a barometer for what would likely be available at the time of the study launch.
Candidate gene-condition pairs for Panel 2 were identified from the NBSTRN RUSP candidate list (20), the ASGCT Clinical Trials Finder, and the list of gene-condition pairs that were reviewed but not approved for Panel 1. Gene-condition pairs were discussed at GPWG meetings, followed by an anonymous vote using the Zoom poll feature to determine inclusion on or exclusion from Panel 2. Gene-condition pairs that received a majority of positive votes were approved for inclusion; group consensus was not required.
Technical and Clinical Feasibility Assessment
To implement these panels in a clinical diagnostic laboratory using DBS, the genes needed further assessment for technical limitations and clinical considerations for variant classification and reporting. As the DBS sample type is not commonly used for genome sequencing, some limitations to testing are expected. The genes were evaluated for technical feasibility, including 15x coverage, homology concerns, special technical requirements for complex common pathogenic variants, and copy number variation calling. If successful gene variant calling was not feasible on the DBS samples, due to any of the technical limitations, the gene was removed from the panel.
Clinical consideration for variant classification and reporting was critical given the presymptomatic/asymptomatic cohort. The reportability of variants was determined using a combination of factors, including known modes of inheritance, mechanisms of disease, expert internal data, and published data. Reporting guidelines were created by the laboratory for each gene based on this information so that variant and clinical scientists could consistently and quickly select reportable variants, which is crucial to maintain turnaround times and consistency across all reports for all individuals enrolled. Special consideration was given to genes with both autosomal dominant and recessive phenotypes to limit over-reporting of carrier status in accordance with the tenants of NBS.
Reported Variants
Reporting guidelines limit reportable variants for Panel 1 and Panel 2 to pathogenic (P) and likely pathogenic (LP) variants. In addition, single heterozygous variants in autosomal recessive disorder genes (i.e., carriers) are not routinely reported. For gene-condition pairs with autosomal recessive inheritance, variants will only be reported if two P/LP variants are identified, to avoid returning results consistent with carrier status. Single heterozygous P/LP variants may be reported for genes associated with both autosomal dominant and autosomal recessive phenotypes (e.g., ABCC8). Thus, carrier status may be inadvertently detected. X-linked P/LP variants will be returned in females only for those conditions where heterozygous females are expected to manifest symptoms (e.g., COL4A5—Alport syndrome). While the GPWG felt it was acceptable to report only specific early-onset variants with a gene, this is not feasible for the laboratory. All P/LP variants in a gene will be returned if the gene is included. The only individual variant the GPWG approved was RET:c.2753T>C (p.M918T). Since the lab cannot feasibly return just this single variant, we opted to return all P/LP variants in the RET gene rather than leave RET off Panel 1.
Initial Panel Evaluation
We counted the number of gene-condition pairs on each panel and determined the median and range of ASQM scores. To determine if Panel 1 was truly “RUSP like,” Wilcoxon tests were performed to compare Panel 1 RUSP gene-condition pairs to Panel 1 non-RUSP gene-condition pairs.
Inter-Study Gene List Comparison
A similar newborn sequencing study in New York, the GUARDIAN study, partnered with the same commercial laboratory as Early Check to perform GS. GUARDIAN also offered two panels to parents and launched with a panel of 155 gene-condition pairs on Panel 1 (https://guardian-study.org/conditions/screened/). We compared the overlap and ASQM scores of Early Check Panel 1 and GUARDIAN Panel 1. Wilcoxon tests were performed to compare Early Check Panel 1 to GUARDIAN Panel 1. We also compared the Early Check list to a published list of 55 consensus genes that were included in six newborn sequencing studies (6).
RESULTS
Populating Early Check Panel 1
The GPWG identified 87 gene-condition pairs associated with core RUSP conditions. This “molecular RUSP” included 7 genes associated with SCID, 31 genes associated with autosomal recessive prelingual nonsyndromic hearing loss, and 9 genes associated with congenital hypothyroidism (Table S1).
The GPWG next reviewed and discussed all gene-condition pairs with ASQM scores ≥13 for Panel 1. The ASQM was previously used to score > 800 gene-condition pairs (Figure 2) enriched for pediatric onset (16, 21). Of these, 137 received a score ≥ 13. After the removal of molecular RUSP gene-condition pairs, the GPWG had 95 gene-condition pairs to review. “Early rule out” excluded 17 due to age of action clearly outside ≤ 2 years (e.g., BRCA1—hereditary breast and ovarian cancer). Of the 78 gene-condition pairs scoring ≥ 13 that were curated and reviewed by the GPWG, 69 (88%) were approved for inclusion (Table S1, Table S5). An additional 42 gene-condition pairs that scored between 9 and 12 and were either nominated or belonged to a group of conditions with one scoring ≥ 13 (e.g., urea cycle disorders) were reviewed, and 33 (79%) were approved for inclusion (Table S1, Table S5). The most-common reason for exclusion was a typical age of action above 2 years (e.g., FBN1—Marfan syndrome). Additional reasons for exclusion are shown in Table 1. In total, the GPWG approved 189 gene-condition pairs for Panel 1 (87 molecular RUSP + 102 non-RUSP). Exemplar gene-condition pairs, their ASQM scores, and reasons for excluding those not approved are listed in Table 1. Additional gene-condition pairs continue to be nominated and evaluated by the GPWG throughout the course of the study and may be added to Panel 1 at 6- to 12-month intervals.
Figure 2:
The distribution of ASQM scores and age of intervention implementation of previously scored gene-condition pairs (n = 843). Target gene-condition pairs for Early Check Panel 1 had ASQM scores ≥ 9 and an age of intervention implementation of ≤ 2 years.
Table 1.
Examples of gene-condition pairs that were and were not approved for Panel 1 by the GPWG.
| Gene | Condition | Typical Age of Action ≤ 2 Years | ASQM Score | |||||
|---|---|---|---|---|---|---|---|---|
| Severity | Likelihood | Efficacy | Acceptability | Knowledge | Total | |||
| Approved for Panel 1 (Exemplars) | ||||||||
| ALDOB | Fructose intolerance | Yes | 2 | 3 | 3 | 3 | 3 | 14 |
| ALPL | Hypophosphatasia | Yes | 2 | 3 | 2 | 2 | 2 | 11 |
| COL4A5 | Alport syndrome | Yes | 2 | 3 | 2 | 3 | 3 | 13 |
| CYBB | Chronic granulomatous disease, X-linked | Yes | 2 | 3 | 3 | 2 | 2 | 12 |
| CYP27A1 | Cerebrotendinous xanthomatosis | Yes | 2 | 3 | 3 | 3 | 2 | 13 |
| F8 | Hemophilia A | Yes | 2 | 3 | 3 | 3 | 3 | 14 |
| G6PC | Glycogen storage disease 1A | Yes | 2 | 3 | 3 | 2 | 3 | 13 |
| G6PD | G6PD deficiency | Yes | 1 | 2 | 3 | 3 | 3 | 12 |
| GJB2 | Deafness, autosomal recessive 1A | Yes | 1 | 3 | 3 | 3 | 3 | 13 |
| IL2RG | SCID, X-linked | Yes | 3 | 3 | 3 | 3 | 3 | 15 |
| INS | Diabetes mellitus, permanent neonatal | Yes | 2 | 3 | 3 | 3 | 2 | 13 |
| KCNQ1 | Jervell and Lange-Nielsen syndrome | Yes | 3 | 3 | 2 | 2 | 3 | 13 |
| RB1 | Retinoblastoma | Yes | 2 | 3 | 3 | 2 | 3 | 13 |
| RET | Multiple endocrine neoplasia type 2B | Yes | 2 | 3 | 3 | 2 | 3 | 13 |
| Not Approved for Panel 1 (Exemplars) | ||||||||
| APOB | Familial hypercholesterolemia | No | 2 | 3 | 3 | 2 | 3 | 13f |
| CD3E | Immunodeficiency 18 (SCID) | Yes | 3 | 3 | 3 | 2 | 1 | 12b |
| COL11A2 | Deafness, autosomal recessive 53 | Yes | 1 | 3 | 3 | 3 | 1 | 11b,e |
| FBN1 | Marfan syndrome | No | 2 | 3 | 2 | 3 | 3 | 13f |
| GCK | MODY, type II | No | 0 | 3 | 0 | 0 | 3 | 6a,d,f |
| GCH1 | Dystonia, DOPA-responsive | No | 1 | 3 | 3 | 3 | 3 | 13d,f |
| GLA | Fabry disease | No | 2 | 3 | 2 | 2 | 3 | 12f |
| PLAU | Quebec platelet disorder | Yes | 1 | 3 | 3 | 3 | 3 | 13c,e |
| PROP1 | Pituitary hormone deficiency, combined, 2 | No | 1 | 3 | 3 | 3 | 3 | 13d,f |
| SLC25A15 | HHH syndrome | Yes | 1 | 3 | 2 | 2 | 2 | 10a |
| TTPA | Ataxia with isolated vitamin E deficiency | Yes | 1 | 3 | 3 | 3 | 3 | 13a |
Reasons for excluding a gene-condition pair included the majority of the GPWG concluding that a) there was insufficient evidence to support genotyping in the absence of a clinical phenotype
the gene accounted for an exceedingly small portion of the associated phenotype
pathogenic variants in the gene are rare outside of their founder populations
the gene-condition was more appropriate for inclusion in a diagnostic panel
moderate, limited, or disputed ClinGen-assessed gene-disease validity
the typical age of action was above 2 years.
Populating Early Check Panel 2
The GPWG curated evidence summaries for 55 candidate gene-condition pairs for consideration in Panel 2; 32 (58%) were approved for inclusion (Table S2, Table S6). The most-common reasons for exclusion were lack of approved therapies or qualifying clinical trials (e.g., HEXA—Tay-Sachs disease). Table 2 includes exemplar gene-condition pairs, their ASQM scores, and reasons for exclusion in cases where approval was not granted. Additional gene-condition pairs will be evaluated by the GPWG throughout the course of the study and may be added to Panel 2 at 6- to 12-month intervals.
Table 2.
Examples of gene-condition pairs that were and were not approved for Panel 2 by the GPWG.
| Gene | Condition | Reason Approved/Not Approved | ASQM Score | |||||
|---|---|---|---|---|---|---|---|---|
| Severity | Likelihood | Efficacy | Acceptability | Knowledge | Total | |||
| Approved for Panel 2 (Exemplars) | ||||||||
| ASPA | Canavan disease | b | 3 | 3 | 0 | 0 | 3 | 9 |
| DMD | Duchenne muscular dystrophy | a | 1 | 3 | 1 | 2 | 2 | 9 |
| GUSB | MPS VII | a | 2 | 3 | 2 | 1 | 1 | 9 |
| MECP2 | Rett syndrome | a | 2 | 3 | 1 | 2 | 1 | 9 |
| RPE65 | Leber congenital amaurosis 2 | a | 1 | 3 | 2 | 2 | 2 | 10 |
| SCN1A | Dravet syndrome | a, b | 2 | 2 | 2 | 2 | 2 | 10 |
| SLC6A8 | Creatine transporter deficiency | a | 1 | 3 | 0 | 0 | 2 | 6 |
| TTPA | Ataxia with isolated vitamin E deficiency | a | 1 | 3 | 3 | 3 | 3 | 13 |
| Not Approved for Panel 2 (Exemplars) | ||||||||
| ARSA | Metachromatic leukodystrophy | c, d, g | 2 | 3 | 1 | 1 | 2 | 9 |
| CLN3 | Ceroid lipofuscinosis, neuronal 3 | c, e, g | 2 | 3 | 0 | 0 | 3 | 8 |
| CTSA | Galactosialidosis | c, f | 3 | 2 | 0 | 0 | 1 | 6 |
| ETHE1 | Ethylmalonic encephalopathy | d | 3 | 3 | 1 | 3 | 1 | 11 |
| GNS | MPS IIID | c, f, g | 2 | 3 | 0 | 0 | 2 | 7 |
| TBX1 | DiGeorge syndrome | d, e | 2 | 2 | 2 | 3 | 1 | 10 |
Reasons for inclusion were the majority of the GPWG concluded that a) approved therapies were available and evidence was sufficient to include in Panel 2 or b) active, qualifying clinical trials were identified. Reasons for exclusion included that c) trials were concluded but therapies were not yet approved in the United States, d) approved therapies lacked sufficient evidence for inclusion in Panel 2, e) qualifying clinical trials did not enroll children ≤ 2 years, and f) therapeutics were in pre-clinical stages. On watch list (g), as new therapies/clinical trials in the United States are expected soon.
Technical and Clinical Feasibility Assessment
Upon review by the partnering clinical diagnostic laboratory, 14 gene-condition pairs had to be removed from Panel 1 or Panel 2 due to high carrier frequency reporting or technical limitations. Eleven gene-condition pairs were removed from Panel 1, including four gene-conditions pairs on the “molecular RUSP” (Table S3). Three gene-condition pairs were removed from Panel 2 (Table S4). The most-common reasons for removal were problems with pseudogenes (e.g., CYP21A2—Adrenal hyperplasia, congenital, due to 21-hydroxylase deficiency), variant detection not being currently validated on GS pipeline with DBS samples (e.g., HBA1—Alpha-thalassemia), and high carrier frequency reporting (e.g., screening for autosomal dominant hearing loss associated with GJB2 would also detect carriers of the more common autosomal recessive form).
Initial Panel Evaluation
At the time of launch, Panel 1 (Table S1) included 178 gene-condition pairs (169 unique genes), including 83 “molecular RUSP” gene-condition pairs, and Panel 2 (Table S2) included 29 gene-condition pairs (29 unique genes). The median ASQM score of the molecular RUSP gene-condition pair set was 12 (range 4 to 15) with 79% (n = 69) scoring ≥ 12 and 48% (n = 42) scoring ≥ 13 (Figure 3). The gene-condition pair scoring a 4 is MCCC2—3-methylcrotonyl-CoA carboxylase deficiency (3-MCC deficiency), which has low penetrance and usually does not require treatment but was included in Panel 1, as it can be detected on state NBS. The median ASQM score of non-RUSP gene-condition pairs in Panel 1 was 13 (range 9 to 15), with 86% (n = 82) scoring ≥ 12 and 68% (n = 65) scoring ≥ 13 (Figure 3). Comparing ASQM scores between RUSP and non-RUSP gene-condition pairs there was a small effect size (r=0.21) but a significant difference (p < 0.01) with the non-RUSP gene-condition pairs scoring higher.
Figure 3:
The distribution of ASQM scores for all gene-condition pairs included in Panel 1, RUSP conditions in Panel 1 and non-RUSP conditions in Panel 1.
ASQM scores of gene-condition pairs in Panel 2 range from 6 to 14, with 76% (n = 22) scoring ≤ 11 (Figure 4). The median ASQM score for Panel 2 gene-condition pairs was 10.
Figure 4:
The distribution of ASQM scores for gene-condition pairs included in Panel 2.
Gene List Comparison
Less than half (49%) of gene-condition pairs included in the initial Early Check Panel 1 are also included in GUARDIAN Panel 1. The Early Check and GUARDIAN panels have 87 overlapping gene-condition pairs (Table S7). GUARDIAN includes 68 gene-condition pairs not included in Early Check, and Early Check includes 91 gene-condition pairs not included in GUARDIAN. Most of the GUARDIAN gene-condition pairs had ASQM scores (144/155). The median ASQM scores of gene-condition pairs in GUARDIAN Panel 1 was 12 (range 1 to 15), with 61% (n = 88) scoring ≥ 12 and 50% (n = 57) scoring ≥ 13. Seventeen (12%) gene-condition pairs in GUARDIAN Panel 1 had ASQM scores of < 9. Comparing ASQM scores between Early Check and GUARDIAN there was a small effect size (r=0.24) but a significant difference (p < 0.0001).
From the recently published consensus list of 55 genes (6), 47 are in Early Check Panel 1 and three (SLC25A15, SLC25A20, SLC2A1) are in Early Check Panel 2. The GPWG was split on whether to include SLC25A15—HHH syndrome—in Panel 1 and ultimately decided to include it in Panel 2 because available treatment does not impact the neurocognitive part of the phenotype. SLC25A20—carnitine-acylcarnitine translocase deficiency—was put in Panel 2 because most affected individuals present within the first 48 hours of life, before GS results would be available. There is less evidence about the treatment in the few patients who present later and would receive sequencing results in time to intervene before symptom onset. Finally, SLC2A1—GLUT1 deficiency syndrome—was put in Panel 2 because available evidence of the intervention demonstrated high efficacy for improvement in epilepsy but less evidence for improved cognition (22). Five genes with ASQM scores of 12 were not included in either Early Check panel at the time of launch, but they were subsequently approved by the GPWG for the next iteration of Panel 1. A majority also voted to move SLC2A1—GLUT1 deficiency syndrome from Panel 2 to Panel 1 for the next iteration of the panels.
DISCUSSION
The integration of genomic sequencing into standard public health NBS is increasingly likely as the technologies advance and become more cost effective (3, 23), which could lead to a significant expansion in the number of conditions screened with a single test. Simultaneously, the development of disease-modifying gene and cell therapies is underway for many genetic conditions detectable through sequencing (24). Currently, some conditions otherwise meeting Wilson and Jungner criteria for NBS inclusion cannot be screened due to the absence of available biomarkers. Advancements in NBS genomic sequencing offer revolutionary capabilities to detect a large number of genetic disorders in newborns with a single test, enabling timely initiation of treatment (4). However, critical questions remain about the feasibility, acceptability, and utility of NBS genomic sequencing.
In the next few years, over a dozen groundbreaking research studies, including the Early Check study, will conduct pilot studies of newborn sequencing, providing valuable insights regarding the potential for future implementation in public health (5). While the potential benefits of newborn sequencing are considerable, they must be weighed against numerous ethical, legal, and social implications (25). For example, genomic sequencing has reduced sensitivity compared to standard newborn screening methods (e.g., tandem mass spectrometry) (26). Variants of uncertain significance (VUS), which most current newborn sequencing studies do not return, are one reason for the reduced sensitivity of genomic sequencing. Due to underrepresentation in genomic research, VUS are more common in individuals who are not European white (27). Thus, widespread implementation of genomic newborn screening has the potential to increase health care disparities.
Some critical implications hinge on decisions regarding which genes are included in screening. The stakes are high, with many experts cautioning about potential adverse consequences to NBS; one expert warned that “adding genomic sequencing to NBS might jeopardize all of NBS, which has been and continues to be so beneficial for thousands of children and their families throughout the world” (28). Transparency and rigor in gene selection will help ensure that NBS using genomic sequencing is acceptable to the population, another consideration of the Wilson and Jungner criteria (29). Ultimately, the genes included may also weigh into ethical decisions around the need for parental consent should genomic screening be incorporated into public health NBS. Prior research has shown that many parents are in favor of a consented approach that allows some degree of choice (30, 31).
Decisions made by current research studies are expected to shape future gene panels and decision frameworks for potential public health implementation. However, investigators conducting pilot studies on newborn sequencing are individually devising methods to determine which genes and conditions to include. Consequently, there is a notable lack of overlap in gene/condition lists among publicly available resources, with large percentages of genes differing between studies (6). In comparing the initial Early Check and GUARDIAN gene lists—two studies partnering with the same clinical sequencing laboratory—only about half of the included genes overlapped. An explanation for the difference is not immediately obvious apart from the fact that the studies used different methods to populate the panels, and/or that the initial priority gene-conditions may have differed (which would suggest increased concordance over time). The methods GUARDIAN used to populate their panels have not yet been published, limiting further insight into explanations for the differences between the two study gene lists. Five other studies have made public their approaches to develop their panels, which included between 318 and 888 genes (6, 32–34). These five studies developed inclusion criteria with a treatability component but none of these studies utilized the ASQM or another semi-quantitative method to assess actionability and determine what genes to include. Collaborative efforts among newborn sequencing studies to establish shared criteria or a uniform list of gene-condition pairs for assessment across all studies could enhance cross-study comparisons. The formation of the International Consortium on Newborn Sequencing in 2022 aims to foster cross-study collaboration (35). The comparison of frameworks and gene lists has been identified as a high-priority activity, and newer studies are currently in the process of developing their gene lists.
The Early Check process offers a reproducible methodology for selecting gene-condition pairs for NBS sequencing pilot studies. Yet, there are limitations to Early Check’s approach of populating newborn sequencing panels. It is a labor-intensive process that requires manual review and updates as new therapies are approved. There is also a component of subjectivity that may differ between working groups composed of different members with varying viewpoints. Ideally in a public health setting, a broader group of interested parties would be involved in decision making including patients, parents, public health professionals, and individuals from minoritized populations. The Early Check team plans to assess the values and priorities of these invested parties in future research.
Early Check has systematically identified a highly actionable set of 178 gene-condition pairs that are appropriate to screen in newborns. ASQM scores of non-RUSP conditions have a similar distribution to ASQM scores of molecular RUSP conditions, indicating that these gene-condition pairs are in fact RUSP-like. Significantly, the Early Check frameworks provide a method for continuous assessment of gene-condition pairs, informing their future integration into population-level screening. Because ASQM scores are dynamic, with increasing actionability as new therapies emerge, an ongoing process is crucial for re-evaluating gene-condition pairs over time. Early Check plans to regularly re-evaluate gene-condition pairs in Panel 2, considering their candidacy for Panel 1 as new therapies become available. There is also the possibility of decreasing actionability if newly approved therapies prove to be less effective than clinical trials or early clinical data suggested. Through newborn screening research we may gain additional data on the effectiveness of interventions in pre-symptomatic children, which are difficult to gather without newborn screening. Early Check was developed back in 2018 expressly for that purpose (7). These data may also lead to changes in ASQM scores over time.
Prior to potential adoption in public health settings, additional important considerations, such as the cost and availability of interventions, and the cost-effectiveness of screening and follow-up need thorough assessment. Data on the true prevalence of positive results in unselected populations, as opposed to prevalence estimates from clinically diagnosed individuals, will help inform workforce needs. Furthermore, other invested parties, including parents of healthy newborns, parents of newborns with genetic conditions, public health professionals, primary care providers, and specialists, should play a role in establishing a consensus on the criteria for genes screened for by sequencing as the field progresses.
Supplementary Material
Acknowledgments
We acknowledge the critical support of the Early Check collaborating organizations: NCSLPH and UNC. The findings and conclusions in this publication are those of the authors and do not necessarily represent the views of the North Carolina Department of Health and Human Services, Division of Public Health. We acknowledge Don Bailey for acting as a senior advisor to the project, Emily Cheves for project management assistance, Marcia Underwood for graphics design, and Angela You Gwaltney for the statistical gene list comparisons.
Funding
This work was initially supported by a planning grant from Janssen. Study implementation is supported by Breakthrough T1D (formerly JDRF International), The Leona M. and Harry B. Helmsley Charitable Trust, Travere Therapeutics, and Orchard Therapeutics. Genome sequencing reagents are provided by Illumina, and genome sequencing is provided by GeneDx. Early Check was also supported by grants from the John Merck Fund and the National Center for Advancing Translational Sciences of the National Institutes of Health (Award #UL1TR002489).
Footnotes
Conflict of Interest: RSZ, SFS, AB, KGL, KGM, CK, KSH, and PK were employees of GeneDx, LLC at the time of this study. The remaining authors have declared that no competing interests exist
Ethics Declaration: The Early Check project has been reviewed and approved by the University of North Carolina at Chapel Hill IRB (#18–0009).
Supplementary Material: Gene-condition pairs on Early Check Panel 1 (Table S1) and Early Check Panel 2 (Table S2). Gene-condition pairs approved but removed from Panel 1 (Table S3) and Panel 2 (Table S4) due to technical limitations. Gene-conditions pairs not approved for Panel 1 (Table S5) and Panel 2 (Table S6). Gene-condition pairs on GUARDIAN Panel 1 (Table S7). Example evidence summary for SCN1A (supplemental file 1).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Data Availability
All data are provided in supplementary files.
References
- 1.Berg JS, Powell CM. Potential Uses and Inherent Challenges of Using Genome-Scale Sequencing to Augment Current Newborn Screening. Cold Spring Harb Perspect Med. 2015;5(12). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Berg JS, Agrawal PB, Bailey DB Jr., Beggs AH, Brenner SE, Brower AM, et al. Newborn Sequencing in Genomic Medicine and Public Health. Pediatrics. 2017;139(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bick D, Ahmed A, Deen D, Ferlini A, Garnier N, Kasperaviciute D, et al. Newborn screening by genomic sequencing: opportunities and challenges. Int J Neonatal Screen. 2022;8(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Woerner AC, Gallagher RC, Vockley J, Adhikari AN. The use of whole genome and exome sequencing for newborn screening: challenges and opportunities for population health. Front Pediatr. 2021;9:663752. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Stark Z, Scott RH. Genomic newborn screening for rare diseases. Nat Rev Genet. 2023;24(11):755–66. [DOI] [PubMed] [Google Scholar]
- 6.Downie L, Bouffler SE, Amor DJ, Christodoulou J, Yeung A, Horton AE, et al. Gene selection for genomic newborn screening: Moving toward consensus? Genet Med. 2024;26(5):101077. [DOI] [PubMed] [Google Scholar]
- 7.Bailey DB Jr., Gehtland LM, Lewis MA, Peay H, Raspa M, Shone SM, et al. Early Check: translational science at the intersection of public health and newborn screening. BMC Pediatr. 2019;19(1):238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kucera KS, Taylor JL, Robles VR, Clinard K, Migliore B, Boyea BL, et al. A Voluntary Statewide Newborn Screening Pilot for Spinal Muscular Atrophy: Results from Early Check. Int J Neonatal Screen. 2021;7(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kucera KS, Boyea BL, Migliore B, Potter SN, Robles VR, Kutsa O, et al. Two years of newborn screening for Duchenne muscular dystrophy as a part of the statewide Early Check research program in North Carolina. Genet Med. 2024;26(1):101009. [DOI] [PubMed] [Google Scholar]
- 10.Okoniewski KC, Wheeler AC, Lee S, Boyea B, Raspa M, Taylor JL, et al. Early Identification of Fragile X Syndrome through Expanded Newborn Screening. Brain Sci. 2019;9(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Paquin RS, Lewis MA, Harper BA, Moultrie RR, Gwaltney A, Gehtland LM, et al. Outreach to new mothers through direct mail and email: recruitment in the Early Check research study. Clin Transl Sci. 2021;14(3):880–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sharp SA, Rich SS, Wood AR, Jones SE, Beaumont RN, Harrison JW, et al. Development and standardization of an improved type 1 diabetes genetic risk score for use in newborn screening and incident diagnosis. Diabetes Care. 2019;42(2):200–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kemper AR, Green NS, Calonge N, Lam WK, Comeau AM, Goldenberg AJ, et al. Decision-making process for conditions nominated to the recommended uniform screening panel: statement of the US Department of Health and Human Services Secretary’s Advisory Committee on Heritable Disorders in Newborns and Children. Genet Med. 2014;16(2):183–7. [DOI] [PubMed] [Google Scholar]
- 14.Roman TS, Crowley SB, Roche MI, Foreman AKM, O’Daniel JM, Seifert BA, et al. Genomic sequencing for newborn screening: results of the NC NEXUS Project. Am J Hum Genet. 2020;107(4):596–611. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hunter JE, Jenkins CL, Bulkley JE, Gilmore MJ, Lee K, Pak CM, et al. ClinGen’s Pediatric Actionability Working Group: Clinical actionability of secondary findings from genome-scale sequencing in children and adolescents. Genet Med. 2022;24(6):1328–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Milko LV, O’Daniel JM, DeCristo DM, Crowley SB, Foreman AKM, Wallace KE, et al. An age-based framework for evaluating genome-scale sequencing results in newborn screening. J Pediatr. 2019;209:68–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Health Resources & Services Administration. Recommended uniform screening panel 2023. [Available from: https://www.hrsa.gov/advisory-committees/heritable-disorders/rusp.
- 18.Strande NT, Riggs ER, Buchanan AH, Ceyhan-Birsoy O, DiStefano M, Dwight SS, et al. Evaluating the Clinical Validity of Gene-Disease Associations: An Evidence-Based Framework Developed by the Clinical Genome Resource. Am J Hum Genet. 2017;100(6):895–906. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.American Academy of Pediatrics. Revised American Thyroid Association guidelines for the management of medullary thyroid carcinoma. Pediatrics. 2018;142(6). [DOI] [PubMed] [Google Scholar]
- 20.Chan K, Hu Z, Bush LW, Cope H, Holm IA, Kingsmore SF, et al. NBSTRN tools to advance newborn screening research and support newborn screening stakeholders. Int J Neonatal Screen. 2023;9(4). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.DeCristo DM, Milko LV, O’Daniel JM, Foreman AKM, Mollison LF, Powell BC, et al. Actionability of commercial laboratory sequencing panels for newborn screening and the importance of transparency for parental decision-making. Genome Med. 2021;13(1):50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Schwantje M, Verhagen LM, van Hasselt PM, Fuchs SA. Glucose transporter type 1 deficiency syndrome and the ketogenic diet. J Inherit Metab Dis. 2020;43(2):216–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Remec ZI, Trebusak Podkrajsek K, Repic Lampret B, Kovac J, Groselj U, Tesovnik T, et al. Next-generation sequencing in newborn screening: a review of current state. Front Genet. 2021;12:662254. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Shahryari A, Saghaeian Jazi M, Mohammadi S, Razavi Nikoo H, Nazari Z, Hosseini ES, et al. Development and clinical translation of approved gene therapy products for genetic disorders. Front Genet. 2019;10:868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lantos JD. Ethical and psychosocial issues in Whole Genome Sequencing (WGS) for newborns. Pediatrics. 2019;143(Suppl 1):S1–S5. [DOI] [PubMed] [Google Scholar]
- 26.Bodian DL, Klein E, Iyer RK, Wong WS, Kothiyal P, Stauffer D, et al. Utility of whole-genome sequencing for detection of newborn screening disorders in a population cohort of 1,696 neonates. Genet Med. 2016;18(3):221–30. [DOI] [PubMed] [Google Scholar]
- 27.Chen E, Facio FM, Aradhya KW, Rojahn S, Hatchell KE, Aguilar S, et al. Rates and Classification of Variants of Uncertain Significance in Hereditary Disease Genetic Testing. JAMA Netw Open. 2023;6(10):e2339571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Levy HL. Ethical and psychosocial implications of genomic newborn screening. Int J Neonatal Screen. 2021;7(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Andermann A, Blancquaert I, Beauchamp S, Dery V. Revisiting Wilson and Jungner in the genomic age: a review of screening criteria over the past 40 years. Bull World Health Organ. 2008;86(4):317–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Armstrong B, Christensen KD, Genetti CA, Parad RB, Robinson JO, Blout Zawatsky CL, et al. Parental Attitudes Toward Standard Newborn Screening and Newborn Genomic Sequencing: Findings From the BabySeq Study. Front Genet. 2022;13:867371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Joseph G, Chen F, Harris-Wai J, Puck JM, Young C, Koenig BA. Parental Views on Expanded Newborn Screening Using Whole-Genome Sequencing. Pediatrics. 2016;137 Suppl 1(Suppl 1):S36–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ceyhan-Birsoy O, Machini K, Lebo MS, Yu TW, Agrawal PB, Parad RB, et al. A curated gene list for reporting results of newborn genomic sequencing. Genet Med. 2017;19(7):809–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kingsmore SF, Smith LD, Kunard CM, Bainbridge M, Batalov S, Benson W, et al. A genome sequencing system for universal newborn screening, diagnosis, and precision medicine for severe genetic diseases. Am J Hum Genet. 2022;109(9):1605–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.England G The Generation Study Protocol 2023. [Available from: https://www.genomicsengland.co.uk/assets/images/Generation-Study-Protocol-V4.0.pdf.
- 35.International Consortium on Newborn Sequencing (ICoNS). ICoNS home; page 2023 [Available from: https://www.iconseq.org/the-consortium. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data are provided in supplementary files.