At-Risk Genomic Findings for Pediatric-Onset Disorders From Genome Sequencing vs Medically Actionable Gene Panel in Proactive Screening of Newborns and Children

Key Points

Question

Is there a clinical value of screening ostensibly healthy newborns and children with genome sequencing in comparison with a gene panel for medically actionable pediatric conditions?

Findings

In this case series study, genome sequencing uncovered potential pediatric-onset diagnoses in 8.2% of apparently healthy children, with 46.8% of findings associated with high-penetrance conditions. In contrast, only 2.1% of children screened with a panel of 268 genes for medically actionable pediatric conditions were found to be at risk for developing pediatric-onset disease, a significant difference.

Meaning

This study suggests that, when compared with a gene panel, genome sequencing uncovered more potential pediatric-onset diagnoses among apparently healthy children.

Abstract

Importance

Although the clinical utility of genome sequencing for critically ill children is well recognized, its utility for proactive pediatric screening is not well explored.

Objective

To evaluate molecular findings from screening ostensibly healthy children with genome sequencing compared with a gene panel for medically actionable pediatric conditions.

Design, Setting, and Participants

This case series study was conducted among consecutive, apparently healthy children undergoing proactive genetic screening for pediatric disorders by genome sequencing (n = 562) or an exome-based panel of 268 genes (n = 606) from March 1, 2018, through July 31, 2022.

Exposures

Genetic screening for pediatric-onset disorders using genome sequencing or an exome-based panel of 268 genes.

Main Outcomes and Measures

Molecular findings indicative of genetic disease risk.

Results

Of 562 apparently healthy children (286 girls [50.9%]; median age, 29 days [IQR, 9-117 days]) undergoing screening by genome sequencing, 46 (8.2%; 95% CI, 5.9%-10.5%) were found to be at risk for pediatric-onset disease, including 22 children (3.9%) at risk for high-penetrance disorders. Sequence analysis uncovered molecular diagnoses among 32 individuals (5.7%), while copy number variant analysis uncovered molecular diagnoses among 14 individuals (2.5%), including 4 individuals (0.7%) with chromosome scale abnormalities. Overall, there were 47 molecular diagnoses, with 1 individual receiving 2 diagnoses; of the 47 potential diagnoses, 22 (46.8%) were associated with high-penetrance conditions. Pathogenic variants in medically actionable pediatric genes were found in 6 individuals (1.1%), constituting 12.8% (6 of 47) of all diagnoses. At least 1 pharmacogenomic variant was reported for 89.0% (500 of 562) of the cohort. In contrast, of 606 children (293 girls [48.3%]; median age, 26 days [IQR, 10-67 days]) undergoing gene panel screening, only 13 (2.1%; 95% CI, 1.0%-3.3%) resulted in potential childhood-onset diagnoses, a significantly lower rate than those screened by genome sequencing (P < .001).

Conclusions and Relevance

In this case series study, genome sequencing as a proactive screening approach for children, due to its unrestrictive gene content and technical advantages in comparison with an exome-based gene panel for medically actionable childhood conditions, uncovered a wide range of heterogeneous high-penetrance pediatric conditions that could guide early interventions and medical management.

This case series study evaluates molecular findings from screening ostensibly healthy children with genome sequencing compared with a gene panel for medically actionable pediatric conditions.

Introduction

Genome sequencing (GS) has entered mainstream medicine; however, cost considerations and reimbursement have hindered widespread acceptance. Genome sequencing offers higher sensitivity than exome sequencing (ES) and microarray testing combined by providing an all-inclusive technological solution for identification of single-nucleotide variants, small insertions or deletions, and copy number variants (CNVs) across the genome.^1,2,3,4 Genome sequencing also delivers high coverage of mitochondrial genomes (mitochondrial DNA [mtDNA]), enabling simultaneous detection of low-heteroplasmy variants,^5,6 and is amenable for detection of repeat expansions and balanced chromosomal rearrangements.^7,8

Diagnostic GS yield from large, rare disease cohort studies is 20% to 40% depending on disease subtypes.^{9,10,11,12,13} Although the actual magnitude is still debatable, the diagnostic sensitivity of GS surpasses that of other genetic testing options, including ES.^{11,12,14,15,16,17,18} The clinical utility of rapid GS has been best demonstrated for critically ill pediatric patients, affecting health care management and resulting in better outcomes and significant health care cost savings.^{19,20,21,22,23,24}

Genome sequencing has entered the realm of preventive and precision medicine as an alternative to targeted panels to screen for specific disease risk, expanded carrier testing, pharmacogenomic variants, and polygenic risk scores, for example.^25,26,27,28 With decreasing sequencing costs, GS is becoming a viable option for population-wide screening,^{29,30,31,32,33} with particular interest for newborn screening (NBS).^34,35,36 Preliminary studies suggest that the integration of rapid GS with traditional NBS is feasible,³⁷ and large publicly or privately funded initiatives are exploring the feasibility, clinical utility, and cost-effectiveness of GS for newborns.^37,38,39

Here, we present, to our knowledge, the largest consumer-driven data set from proactive screening by clinical GS of a large cohort of ostensibly healthy children, predominantly newborns, for genomic risk for pediatric-onset disease as opposed to screening strategies that involve a predetermined list of conditions. We evaluate the clinical utility of our GS approach by comparing the rate and type of findings with those observed in a similar-sized pediatric cohort screened by an ES-based gene panel (ESGP) for 268 genes associated with actionable pediatric conditions.

Methods

This study is a retrospective case series study of deidentified data obtained from an out-of-pocket genetic testing service. The use of deidentified data for publication has been approved by the internal policy review board and the Western Institutional Review Board (WIRB), now known as WIRB-Copernicus Group, institutional review board. Written informed consent was obtained from the parents or guardians of all children undergoing testing. This manuscript follows the reporting guideline for case series through delineating eligibility criteria, testing strategies, statistical methods, comparing 2 cohorts, and discussing a data-driven hypothesis with its limitations and future implications.

Healthy Screening Options

A healthy screening was offered as the ViaCord service for families enrolled in cord blood or tissue banking. During the second or third routine trimester obstetrics visit, parents were introduced to cord blood or tissue banking along with 2 optional genetic screening options, GS or ESGP (Genetic Insight Panel). This out-of-pocket screening was initiated by parents or legal guardians, who were counseled by an independent physician from a virtual medical practice (Genome Medical) to facilitate the screening referral and delivery of the results. All families received pretesting and posttesting genetic counseling. Test orders were placed after counseling or after the birth of an infant if the family consented before delivery. Only reportedly healthy children as determined during the pretest consultation were referred for the healthy screening. Specimens accepted included whole blood, saliva, dried blood spots (DBS), and extracted genomic DNA. This report includes consecutive children referred to our laboratory for GS screening (n = 562) from March 1 2018, through July 31, 2022, and for ESGP screening (n = 606) from June 1, 2019, through July 31, 2022.

GS and ESGP Screening Method

The GS screening sequenced more than 22 000 genes and analyzed approximately 6000 genes associated with disease, including sequence and CNV analysis of nuclear genes and mtDNA sequencing. Regions with highly homologous or pseudogene sequences were not analyzed except for bioinformatics-assisted screening of the SMN1 (OMIM 600354) gene deletions. Copy number variant analysis was limited to the nuclear genome to detect copy number and/or allelic genomic imbalances, such as deletions, duplications, aneuploidy, unbalanced translocations or inversions, absence of heterozygosity, uniparental disomy (UPD), and mosaicism (>20% across large genomic regions). The ESGP included sequence and CNV analysis of 268 nuclear genes curated for association with medically actionable childhood-onset conditions (eTable 1 in Supplement 1).

Assays were performed at PerkinElmer Genomics laboratory (Pittsburgh, Pennsylvania). Genomic DNA derived from whole blood, saliva, or DBS was extracted using the Chemagic 360-D instrument and Chemagic DNA CS200 kit (PerkinElmer). For GS, a polymerase chain reaction–free sequence library was prepared using the Bioo Scientific NEXTFlex Rapid XP DNA-seq Library kit (PerkinElmer Applied Genomics) on the Sciclone G3 NGSx liquid handling workstation (PerkinElmer). For the ESGP, a library was prepared using Bioo Scientific NEXTFLEX Rapid XP kit (PerkinElmer Applied Genomics) followed by sequence target enrichment using the Agilent SureSelect Focused Exome sequence capture kit (Agilent). Pooled libraries were sequenced as 2 × 151-bp paired-end reads on NovaSeq 6000 (Illumina Inc) at a mean target coverage of 100× for the ESGP and 40× for GS following standard Illumina protocol.

Demultiplexing and sequence data conversion to FASTQ files were performed using the Illumina bcl2fastq converter (Illumina Inc). A sequence was aligned to the human reference genome (GRCh37), and variant calling was performed using the Illumina DRAGEN Bio-IT Platform (Illumina Inc). Variant annotation was performed using SNPEff,⁴⁰ with further variant filtering performed using internal software.

The variants reviewed had a minimum coverage of 8× and an alternate allele frequency of 20% or more. Besides exonic sequences, GS also included analysis of pathogenic variants in deep intronic regions as curated from public databases.

Copy number variant calling and absence of heterozygosity calling were performed using NxClinical software (BioDiscovery).⁴¹ SMN1 deletion analysis was performed on GS data only using in-house bioinformatics tools based on published literature with modifications.^42,43,44

Variant classification followed American College of Medical Genetics and Genomics (ACMG) and American Molecular pathologist standards and guidelines.^45,46,47 At-risk reporting (positive findings) included genotypes underlying childhood-onset monogenic disease, including pediatric conditions from the ACMG secondary finding list, version 3.1 (ACMG SF)⁴⁸; contiguous gene syndromes; or chromosomal abnormalities. The reports included clinically significant (ie, likely pathogenic or pathogenic) heterozygous variants in genes associated with autosomal dominant conditions; hemizygous, homozygous, or 2 heterozygous likely pathogenic or pathogenic variants in genes associated with X-linked or autosomal recessive disorders; numerical and structural chromosomal abnormalities; and UPD. Only likely pathogenic or pathogenic variants at the heteroplasmy level of 5% or more in the mtDNA were reported. Recessive disorder carrier status was not reported, including SMN1 heterozygous deletion. The GS screening also offered an opt-in consent for preemptive pharmacogenetics that included select Clinical Pharmacogenetics Implementation Consortium Level A and PharmGKB 1A variants in 7 genes with clinical utility (CYP2C9 [OMIM 601130], CYP2D6 [OMIM 124030], CYP2C19 [OMIM 124020], CYP3A5 [OMIM 605325], SLCO1B1 [OMIM 604843], UGT1A1 [OMIM 191740], and TPMT [OMIM 187680]) (eTable 2 in Supplement 1).

Data Stratification

The age at testing statistics was calculated based on the specimen reception date by the laboratory. Individuals were stratified into age groups as follows: neonates (≤1 month), infants (1-12 months), toddlers (13-36 months), childhood (4-11 years), and adolescence (12-18 years). Reported genotypes or disorders were stratified to low- (<20%), moderate- (20%-80%), and high-penetrance (>80%) categories³⁹ based on curated genotype-specific or gene- or locus-specific literature. Grouping of reported conditions into larger disease categories is shown in eTable 3 in Supplement 1.

Statistical Analysis

Categorical variables, such as diagnostic outcome, were summarized using proportions with 95% CIs provided where applicable. Median (IQR) values were calculated for continuous variables, such as age at testing. The Fisher exact test and the χ² calculator⁴⁹ were used to evaluate the significance of the differences observed in diagnostic outcomes or the distribution of stratified data. All P values were from 2-tailed tests and results were deemed statistically significant at P < .01.

Results

A total of 562 apparently healthy children (286 girls [50.9%] and 276 boys [49.1%]; median age at testing, 29 days [IQR, 9-117 days; range, 2 days-15 years]) underwent proactive screening by GS, and 606 individuals (293 girls [48.3%] and 313 boys [51.7%]; median age at testing, 26 days [IQR, 10-67 days; range, 2 days-3 years]) were proactively screened by ESGP (Table 1). Neonates and infants constituted the majority in both cohorts (Table 1).

Table 1. Stratification of Individuals and Positive Reports per Age Category and Sex.

Study participants	GS cohort (n = 562)			ESGP cohort (n = 606)
	Individuals, No. (%)	Positive reportsa		Individuals, No. (%)	Positive reports
		No.	% (95% CI)		No.	% (95% CI)
Whole cohort	562 (100.0)	46	8.2 (5.9-10.5)	606 (100.0)	13	2.1 (1.0-3.3)
Neonates	288 (51.2)	23	8.0 (4.9-11.1)	329 (54.3)	4	1.2 (0.0-2.4)
Infants	203 (36.1)	18	8.9 (5.0-12.8)	268 (44.2)	9	3.4 (1.2-5.5)
Toddlers	39 (6.9)	2	5.1 (0.0-12.1)	8 (1.3)	0	NA
Childhood	26 (4.6)	2	7.7 (0.0-17.9)	1 (.2)	0	NA
Adolescence	6 (1.1)	1	16.7 (0.0-46.5)	0	0	NA
Sex
Female	286 (50.9)	21	7.3 (4.3-10.4)	293 (48.3)	8	2.7 (0.9-4.6)
Male	276 (49.1)	25	9.1 (5.7-12.4)	313 (51.7)	5	1.6 (0.2-3.0)

Abbreviations: ESGP, exome sequencing–based gene panel; GS, genome sequencing; NA, not applicable.

^a The rate of positive findings across different age categories is not statistically significant.

Specimen and Testing

Most specimens (GS, 529 [94.1%] and ESGP, 600 [99.0%]) were DBS (eTable 4 in Supplement 1). Additional testing data, such as the rate of test failures, turnaround time, and sequence coverage, were obtained from GS and ESGP orders registered during the period from July 1, 2019, to July 31, 2022 (eTable 5 in Supplement 1). Only 11 samples, all DBS, failed. The mean turnaround time was 1.5 times longer for GS vs ESGP (56 vs 37 days). The mean (SD) sequencing coverage of target regions was 50 (12) times for GS and 231 (88) times for ESGP. We estimated consumer preference for 2 screening options from the number of orders received over the period of 3 years and weighted this for cost difference. Although ESGP had 1.2 times more orders than GS (eTable 5 in Supplement 1), considering the 3.3 times higher cost of GS,⁵⁰ the weighted consumer preference was 1:2.75 in favor of GS.

Positive GS Findings

Genome sequencing uncovered potential diagnoses for 46 individuals (8.2%; 95% CI, 5.9%-10.5%) involving 47 at-risk genotypes (Table 2). One individual (No. 25) had 2 molecular results (G6PD [OMIM 305900] and COL4A3 [OMIM 120070]) (Table 3).^{51,52,53,54,55,56,57,58,59,60}

Table 2. Summary of Findings in GS Cohort.

Finding	Actual GS cohort findings			Simulated ESGP analysis on GS cohort
	GS findings, No.	% of All GS findings (n = 47)	Cohort frequency, No. (%) (n = 562)	ESGP-simulated findings, No.	% of Respective findings in GS cohorta	Cohort frequency, No. (%) (n = 562)
All positive findings	47	100.0	46 (8.2)b	17	36.2	16 (2.8)b
SNVs
Overall	33	70.2	32 (5.7)b	10	30.3	9 (1.6)b
AD or semidominant	23	48.9	23 (4.1)	3	13.0	3 (0.5)
AR	5	10.6	5 (0.9)	3	60.0	3 (0.5)
XL	4	8.5	4 (0.7)	4	100.0	4 (0.7)
mtDNA	1	2.1	1 (0.2)	0	0	0
CNVs
Overall	10	21.3	10 (1.8)	3	30.0	3 (0.5)
Recurrent	6	12.8	6 (1.1)	2	33.3	2 (0.4)
Other large	2	4.3	2 (0.4)	1	50.0	1 (0.2)
Intragenic	2	4.3	2 (0.4)	0	0	0
Chromosome scale
Overall	4	8.5	4 (0.7)	4	100.0	4 (0.7)
Structural	2	4.3	2 (0.4)	2	100.0	2 (0.4)
Numerical	1	2.1	1 (0.2)	1	100.0	1 (0.2)
UPD	1	2.1	1 (0.2)	1	100.0	1 (0.2)
ACMG secondary findings
Overall	6	12.8	6 (1.1)	4	66.7	4 (0.7)
By penetrance
High	22	46.8	22 (3.9)	5	22.7	5 (0.9)
Moderate	10	21.3	9 (1.6)c	6	60.0	5 (0.9)c
Low	15	31.9	15 (2.7)	6	40.0	6 (1.1)
Pharmacogenomic findings, No. (% of all cohort) (n = 562)	NA	NA	NA	NA	NA	NA
At least 1 variant	500 (89.0)			0

Abbreviations: ACMG, American College of Medical Genetics and Genomics; AD, autosomal dominant; AR, autosomal recessive; CNV, copy number variant; ESGP, exome sequencing–based gene panel; GS, genome sequencing; mtDNA, mitochondrial genome; NA, not applicable; SNV, single-nucleotide variant; UPD, uniparental disomy; XL, X linked.

^a No. provided in column “GS findings, No.”

^b One individual received 1 positive finding.

^c Two molecular findings of moderate penetrance found in 1 individual.

Table 3. GS Healthy Screening Findings.

Patient No./Age group/Sex	Gene (transcript) for sequence variants or genomic coordinates for CNVs	c., p. for sequence variants or genes of interests for CNVs	Zygosity	Classification	Associated condition	Inheritance	Penetrance	Reported from ESGP?
1/Neonate/M	BTD (NM_001370658.1)a	c.1270G>C, p.(Asp424His)	Hom	P	Biotinidase deficiency	AR	Low	Yes
2/Childhood/M	BTD (NM_001370658.1)a	c.1270G>C, p.(Asp424His)	Hom	P	Biotinidase deficiency	AR	Low	Yes
3/Neonate/F	LDLR (NM_000527.5)a	c.761A>C, p.(Gln254Pro)	Het	P	Hypercholesterolemia	AD	Moderate	Yes
4/Neonate/M	LDLR (NM_000527.5)a	c.939C>A, p.(Cys313*)	Het	P	Hypercholesterolemia	AD	Moderate	Yes
5/Neonate/M	HNF1A (NM_000545.6)a	c.1135C>A, p.(Pro379Thr)	Het	LP	Monogenic diabetes	AD	High	No
6/Infant/F	MYBPC3 (NM_000256.3)a	c.1624G>C, p.(Glu542Gln)	Het	P	Hypertrophic cardiomyopathy	AD	Moderate	No
7/Infant/F	CHD8 (NM_001170629.1)	c.6130C>T, p.(Arg2044*)	Het	LP	ID and autism	AD	High51	No
8/Neonate/M	COL10A1 (NM_000493.3)	c.1616_1617del, p.(Leu539fs)	Het	LP	Metaphyseal chondrodysplasia	AD	High	No
9/Infant/F	COL11A1 (NM_001854.3)	c.1245 + 1G>A, p.(?)	Het	LP	Stickler syndrome	AD	High	No
10/Infant/F	FGFR3 (NM_000142.4)	c.749C>G, p.(Pro250Arg)	Het	P	Muenke syndrome	AD	High	No
11/Infant/M	MITF (NM_000248.3)	c.345 + 1G>A, p.(?)	Het	LP	Waardenburg syndrome	AD	High52	No
12/Neonate/M	PLN (NM_002667.5)	c.116T>G, p.(Leu39*)	Het	P	Cardiomyopathy	AD	Moderate	No
13/Neonate/M	PPM1D (NM_003620.4)	c.1349dup, p.(Leu450fs)	Het	LP	Jansen de Vries syndrome	AD	High53	No
14/Adolescence/F	RPL21 (NM_000982.3)	c.95G>A, p.(Arg32Gln)	Het	LP	Hypotrichosis	AD	High	No
15/Neonate/M	SDHA (NM_004168.4)	c.1534C>T, p.(Arg512*)	Het	P	Paragangliomas	AD	Low	No
16/Infant/M	SLC39A5 (NM_173596.2)	c.1128del, p.(His377fs)	Het	LP	Myopia	AD	High	No
17/Neonate/M	SPTA1 (NM_003126.4)	c.460_462dup, p.(Leu154dup)	Het	P	Hereditary elliptocytosis	AD	Low	No
18/Childhood/M	ABCA4 (NM_000350.3)	c.2588G>C(;)3292C>T, p.(Gly863Ala(;)Arg1098Cys)	Het	LP(;)LP	Retinal dystrophy	AR	High	No
19/Neonate/M	CYP21A2 (NM_000500.9)	c.923dup(;)955C>T, p.(Leu308fs(;)Gln319*)	Het	P(;)P	Congenital adrenal hyperplasia	AR	High	No
20/Infant/M	MEFV (NM_000243.2)	c.2230G>T, p.(Ala744Ser)	Hom	P	Familial Mediterranean fever	AR	High	Yes
21/Neonate/M	MT-TS1 (NC_012920.1)	m.7471dup	7%b	LP	Mitochondrial disease	mtDNA	Low	No
22/Infant/F	G6PD (NM_001042351.3)	c.131C>G, p.(Ala44Th)	Het	LP	G6PD deficiency (favism)	XL	Lowc	Yes
23/Neonate/F	G6PD (NM_001042351.3)	c.[202G>A;376A>G], p.[(Val68Met);(Asn126Asp)]	Het	P	G6PD deficiency (favism)	XL	Lowc	Yes
24/Infant/M	G6PD (NM_001042351.3)	c.563C>T, p.(Ser188Phe)	Hem	P	G6PD deficiency (favism)	XL	Moderatec	Yes
25/Neonate/M	G6PD (NM_001042351.3)	c.[202G>A;376A>G], p.[(Val68Met;Asn126Asp)]	Hem	P	G6PD deficiency (favism)	XL	Moderatec	Yes
	COL4A3 (NM_000091.4)	c.345del, p.(Pro116fs)	Het	P	Alport syndrome	AD/AR	Moderate	Yes
26/Infant/F	TNFRSF13B (NM_012452.3)	c.310T>C, p.(Cys104Arg)	Het	LP	Common variable immunodeficiency	AD	Low54	No
27/Neonate/F	TNFRSF13B (NM_012452.3)	c.310T>C, p.(Cys104Arg)	Het	LP	Common variable immunodeficiency	AD	Low54	No
28/Neonate/F	TNFRSF13B (NM_012452.3)	c.310T>C, p.(Cys104Arg)	Het	LP	Common variable immunodeficiency	AD	Low54	No
29/Neonate/M	SGCE (NM_003919.2)	c.802dup, p.(Ile268fs)	Het	LP	Myoclonic dystonia	AD	Highd	No
30/Toddler/F	SGCE (NM_003919.2)	c.1210_1213del, p.(Asp404fs)	Het	LP	Myoclonic dystonia	AD	Highd	No
31/Infant/M	IFNGR1 (NM_000416.2)	c.523del, p.(Tyr175fs)	Het	P	Immunodeficiency	AD	Low55	No
32/Infant/M	PROKR2 (NM_144773.2)	c.58del, p.(His20fs)	Het	P	Kallmann syndrome	AD	Low	No
33/Toddler/F	seq[GRCh37] (1)(q22q22) chr1:155395998-155538509del	(ASH1L exon 1-5)e	Het	P	Intellectual disability	AD	High56	No
34/Neonate/F	seq[GRch37] (2)(p16.3p16.3) chr2:51128236-51170160del	(NRXN1exon 4-5)e	Het	P	Neurodevelopmental disorder	AD	Moderate57	No
35/Infant/F	seq[GRCh37] del(X)(p22.33p22.33) chrX:323372-1023863del	(SHOX)e	Het	P	Leri-Weill dyschondrosteosis	PD	High	No
36/Infant/M	seq[GRCh37] del(20)(q13.33q13.33) chr20:61798410-62324518del	KCNQ2	Het	P	20q13.33 microdeletion syndrome	AD	High58	Yes
37/Infant/M	seq[GRCh37] dup(7)(q11.23q11.23) chr7:72641431-74148130dup	(LIMK1, ELN)e	Het	P	7q11.23 microduplication syndromef	AD	High	No
38/Infant/M	seq[GRCh37] dup(17)(p12p12) chr17:14089623-15486483dup	(PMP22)e	Het	P	Charcot Marie Tooth diseasef	AD	High	No
39/Neonate/F	seq[GRCh37] del(1)(q21.1q21.2) chr1:146476110-147486609del	(GJA5, CHD1L)e	Het	P	1q21.1 deletion syndromef	AD	Moderate59,60	No
40/Neonate/F	seq[GRCh37] del(16)(p13.11p13.11) chr16:14892404-16364100del	MYH11 (NDE1, NTAN1)e	Het	P	16p13.11 microdeletion syndromef	AD	Low59	Yes
41/Neonate/F	seq[GRCh37] dup(16)(p11.2p11.2) chr16:29449426-30283525dup	(TBX6), PRRT2e	Het	P	16p11.2 duplication syndromef	AD	Moderate59	Yes
42/Infant/M	seq[GRCh37] dup(22)(q11.21q11.21) chr22:18888333-21677032dup	(HIRA, TBX1)e	Het	P	22q11.2 microduplication syndromef	AD	Low59	No
43/Neonate/Fg	seq[GRCh37] der(4)t(4;9)(p16.1;p24.3) chr4:68838-11547487del; chr9:46236-1484376duph	IDUA (LETM1, NSD2, NELFA) DOCK8e	NA	P	Wolf-Hirschhorn Syndrome	NA	High	Yes, Yesh
44/Neonate/F	seq[GRCh37] trp(12)(p13.33p10) mos chr12:186298-34856686 gain	3 Genes in ESGP	NA	P	Pallister-Killian syndrome	NA	High	Yes
45/Infant/M	seq[GRCh37] (8)x3 mos Chromosome 8 gain	9 Genes in ESGP	NA	P	Mosaic trisomy 8	NA	High	Yes
46/Neonate/F	seq[GRCh37] (16)x2 hmz Chromosome 16 isodisomy	12 Genes in ESGP	NA	VUS	Uniparental isodisomy of 16	NA	Low	Yes

Abbreviations: AD, autosomal dominant; AD/AR, semidominant; AR, autosomal recessive; CNV, copy number variant; ESGP, exome sequencing–based gene panel; F, female; GS, genome sequencing; Hem; hemizygous; Het, heterozygous; Hom, homozygous; ID, intellectual disability; LP, likely pathogenic; M, male; mtDNA, mitochondrial genome; NA, not applicable; P, pathogenic; PD, pseudoautosomal; VUS, variant of uncertain significance; XL, X linked.

^a Medically actionable genes included in American College of Medical Genetics and Genomics secondary finding list.

^b Heteroplasmy level.

^c Penetrance differs for males and females.

^d Penetrance high when inherited paternally, but low if maternally.

^e Genes that are not in ESGP are in parentheses.

^f Syndromes due to recurrent CNVs.

^g Two different CNVs found consistent with the presence of unbalanced translocation.

^h Counted as 1 variant in the text.

Sequence analysis uncovered 33 of 47 of all at-risk genotypes (70.2%), while CNV analysis contributed 14 of 47 potential diagnoses (29.8%), including 10 CNVs (4 microdeletion and 4 microduplication syndromes, 2 partial single gene deletions [ASH1L (OMIM 607999) and NRNX1 (OMIM 600565)]) and 4 chromosome scale rearrangements: 1 likely mosaic isochromosome 12p, 1 mosaic trisomy 8, 1 potential derivative chromosome 4 from translocation t(4;9)(p16.1;p24.3), and UPD of chromosome 16. One individual was found to carry pathogenic mtDNA variant at 7% heteroplasmy level. Six individuals had positive findings in ACMG SF. Six children carried recurrent CNVs (Table 2 and Table 3). Only 5 genes had findings in more than 1 individual. These included pathogenic G6PD (OMIM 305900) variants found in 4 individuals, TNFRSF13B (OMIM 604907) at-risk genotype identified in 3 individuals, and variants involving LDLR (OMIM 606945), BTD (OMIM 609019), and SGCE (OMIM 604149) genes were identified in 2 individuals each (Table 3). Stratification of the genotypes by presumed penetrance revealed that almost half the GS findings (46.8% [22 of 47]) were associated with high-penetrance conditions, among a total of 3.9% of the cohort (22 of 562) (Figure, B, and Table 2).

Figure. Genome Sequencing (GS)–Positive Findings Stratified by Penetrance.

A, Genome sequencing findings stratified by penetrance; exome sequencing–based gene panel (ESGP) Yes and ESGP No refer to the potential of the GS findings to be reported from the ESGP screening. B, Individuals with positive GS reports stratified by age group. C, Genome sequencing findings by disease category.

Most GS diagnoses involved 2 disease categories: metabolic disorders and neurodevelopmental disorders, including monogenic and microdeletion or duplication neurodevelopmental syndromes (Figure, C, and Table 3). Each category accounted for 21.3% (10 of 47) of all diagnoses. At least 1 pharmacogenomic variant was returned for 500 individuals (89.0%) (Table 2).

Positive ESGP Findings

Positive ESGP findings were identified for 13 individuals (2.1%; 95% CI, 1.0%-3.3%) (Table 1). Overall, there were 14 potential diagnoses involving 13 different single-nucleotide variants or insertions or deletions in 11 genes (8 autosomal dominant genes, 2 autosomal recessive genes, and 1 X-linked gene reported for 4 individuals; Table 4). One individual (No. 5) received 2 findings. Disease-causing variants in the ACMG genes were found in 5 individuals, with 1 individual (No. 5) receiving 2 ACMG gene results (GAA [OMIM 606800] and PKP2 [OMIM 602861]). High-penetrance conditions were reported for 2 individuals, making up 18.2% of all ESGP findings (2 of 11). Moderate-penetrance conditions were reported for 7 of 11 individuals (63.6%) and low-penetrance conditions were reported for 5 of 11 individuals (45.5% of all ESGP findings). No diagnoses were identified from CNV analysis (Table 4).

Table 4. Healthy Screening Findings From ESGP.

Patient No./Age/Sex	Gene (transcript)	c., p.	Zygosity	Class	Associated condition	Inheritance	Penetrance
1/Neonate/M	RYR1 (NM_000540.2)a	c.1840C>T, p.(Arg614Cys)	Het	P	Malignant hyperthermia	AD	Moderate
2/Infant/F	SCN5A (NM_198056.2)a	c.1127G>A, p.(Arg376His)	Het	P	Long QT syndrome; Brugada syndrome	AD	Moderate
3/Neonate/F	HNF4A (NM_198056.2)a	c.1198C>T, p.(Arg400*)	Het	LP	Monogenic diabetes	AD	Moderate
4/Neonate/F	LDLR (NM_000527.5)a	c.2416dup, p.(Val806fs)	Het	P	Familial hypercholesterolemia	AD	Moderate
5/Infant/F	GAA (NM_000152.3)a	c.[752C>T;761C>T](;)1411_1414del, p.[(Ser251Leu);(Ser254Leu)](;)(Glu471fs)	Het	P(;)P	Pompe disease	AR	High
	PKP2 (NM_004572.3)a	c.369G>A, p.(Trp123*)	Het	P	Arrhythmogenic right ventricular dysplasia	AD	Low
6/Infant/F	COL4A4 (NM_000092.4)	c.2320G>C, p.(Gly774Arg)	Het	LP	Hematuria	AD/AR	Moderate
7/Infant/M	NF1 (NM_000267.3)	c.5206-2_5216dup, p.(Ala1740fs)	Het	LP	NF1 related	AD	High
8/Infant/M	G6PD (NM_001042351.3)	c.563C>T, p.(Ser188Phe)	Hem	P	G6PD deficiency or favism	XL	Moderate
9/Infant/M	G6PD (NM_001042351.3)	c.844G>C, p.(Asp282His)	Hem	LP	G6PD deficiency or favism	XL	Moderate
10/Infant/F	G6PD (NM_001042351.3)	c.202G>A;376A>G, p.(Val68Met);(Asn126Asp)	Het	P	G6PD deficiency or favism	XL	Low
11/Neonate/F	G6PD (NM_001042351.3)	c.202G>A;376A>G, p.(Val68Met);(Asn126Asp)	Het	P	G6PD deficiency or favism	XL	Low
12/Infant/F	TSHR (NM_000369.2)	c.1349G>A, p.(Arg450His)	Het	LP	Hypothyroidism	AD/AR	Low
13/Infant/M	CFTR (NM_000369.2)	c.1210-12T[5](;)1521_1523del, p.(?)(;)(Phe508del)	Het	P(;)P	CFTR related	AR	Low

Abbreviations: AD, autosomal dominant; AD/AR, semidominant; AR, autosomal recessive; ESGP, exome sequencing–based gene panel; F, female; Hem; hemizygous; Het, heterozygous; LP, likely pathogenic; M, male; P, pathogenic; XL, X linked.

^a Denotes genes in the American College of Medical Genetics and Genomics secondary findings list.

To exclude cohort-specific biases in comparing diagnostic outcomes from GS and ESGP proactive screening approaches, we simulated a panel-like analysis of the GS cohort (Table 2). This analysis revealed that only 2.8% children (16 of 562) in the GS cohort would have received at-risk findings by ESGP, closely resembling the rate of findings in the actual ESGP data set. Only 36.2% GS diagnoses (17 of 47) were uncovered by ESGP, and the proportion was even lower for high-penetrance conditions (22.7% [5 of 22]) (Figure, A).

Discussion

The main focus of this case series study was a retrospective analysis to compare the clinical utility of 2 conceptually different newborn sequencing approaches: one approach that is focused on only well-established medically actionable conditions (the actionable disease-centric approach) vs another approach that is an unbiased approach to evaluation of all known disease-causing genes (the genome-open approach). To our knowledge, this is the first study of this scale that provides side-by-side clinical utility comparison for the 2 conceptually different pediatric screening strategies. The participants in both cohorts underwent initial evaluation for genetic screening by the same group of physicians followed by genetic testing conducted in the same laboratory, which reduces ascertainment and analysis-related biases.

Clinical GS screening included analysis of approximately 6000 genes associated with disease; however, only genotypes associated with pediatric-onset conditions were reported. This GS screening found that 8.2% (46 of 562) of ostensibly healthy newborns and children were at risk for childhood-onset disease. Although the rate is apparently high, a similar rate of pediatric genetic disease risk (8.7% [11 of 127]) was reported in a smaller cohort of apparently healthy newborns screened by ES.^39,61,62

Most of the reportable GS findings were associated with incomplete or age-dependent penetrance and variable expressivity, as expected for an apparently healthy population.⁶³ Stratification of the genotypes by presumed penetrance revealed that almost half of the GS findings (46.8% [22 of 47]) were associated with high-penetrance conditions, in a total of 3.9% (22 of 562) of the cohort (Figure, B, and Table 2).

In contrast, the ESGP of 268 genes associated with actionable pediatric diseases identified only 2.1% (13 of 606) children at risk, a significantly lower rate than identified with the GS approach (P < .001). Only 18.2% of all ESGP findings (2 of 11) involved high-penetrance genotypes, as opposed to 46.8% (22 of 47) from the GS data.

The genetic risks uncovered by GS were heterogeneous, with 8.5% of diagnoses (4 of 47) accounted by chromosome alterations, 21.3% by CNVs (10 of 47), and the remaining by single-nucleotide variants and insertions or deletions in a wide range of genes (Table 2 and Table 3). One finding involved a heteroplasmic mtDNA variant. Most genes or variants were observed once, with only a few genes reported in multiple individuals (G6PD in 4 individuals, TNFRSF13B in 3, BTD in 2, LDLR in 2, and SGCE in 2). Genotypes from the ACMG SF were returned for 1.1% (6 of 562) of the cohort, consistent with reports from other large studies.^61,62 Not all genotypes from the ACMG SF (66.7% [4 of 6]) were identified in the simulated data set because not all ACMG SF pediatric genes were included in the current ESGP version. This shortfall is unlikely to significantly influence the rate of positive findings in the ESGP screening because only 12.8% of all GS findings (6 of 47) involved ACMG SF diagnoses. Microduplication or deletion syndromes were identified in 1.4% of individuals (8 of 562) in the GS cohort, with 1.1% (6 of 562) carrying recurrent CNVs. Most of these large CNVs syndromes would be missed by ESGP except for 3 (16p11.2 microduplication and 16p13.1 and 20q13.33 microdeletions). Copy number variants on 16p11.2 and 16p13.1 would be detected by ESGP for purely serendipitous reasons because ESGP includes 1 gene encompassed in each CNV (PRRT2 [OMIM 614386] and MYH11 [OMIM 160745]) due to reasons not associated with these syndromes. Chromosomal abnormalities were found in 4 individuals (0.7% of cohort and 8.5% of all results), and included Wolf-Hirschhorn syndrome, Pallister-Killian syndrome, and mosaic trisomy 8. The absence of prenatal manifestations in these children could explain the incidental nature of these findings. The clinical significance of UPD of chromosome 16 is still under debate because no imprinted disorders have been associated with this chromosome; however, the possibility of hidden trisomy 16 made the finding worth reporting.^64,65,66

Most GS diagnoses involved 2 disease categories: metabolic disorders and neurodevelopmental disorders, including monogenic and microdeletion or duplication neurodevelopmental syndromes (Figure, C, and Table 3). Each category accounted for 21.3% (10 of 47) of all diagnoses. Not all metabolic disorders identified would have been detected by Recommended Universal Screening Panel (RUSP) NBS, except BTD deficiency, adrenal hyperplasia, and G6PD deficiency in some states. A total of 10.6% (5 of 47) of all findings were due to high-penetrance neurodevelopmental conditions, involving syndromes with intellectual disability and/or autism (CHD8 [OMIM 610528], PPM1D [OMIM 605100], ASH1L, KCNQ2 [OMIM 602235], and 7q11.2 duplication syndrome [OMIM 609757]) that would have been missed by the actionable disease-centric approach, although the clinical outcomes of these individuals may benefit from early intervention services.⁶⁷

In fact, many of the GS findings may influence individuals’ health management, including medications, early intervention, disease surveillance, and avoidance of aggravating factors, which could lead to better prognoses and clinical outcomes (eTable 3 in Supplement 1). Given the highly heterogenous and clinically impactful GS findings, ESGP would need to include many more genes, perhaps even full exomes, to match the clinical GS sensitivity. Even then, certain variants, such as heteroplasmic mtDNA variants, mosaic chromosomal abnormalities, and exon-size deletions or duplications, could be missed.⁶⁸ In addition, clinical GS utility should be viewed throughout the lifetime of the individual. Genomic data can be revisited in the future for additional needs, including reanalysis, carrier testing, and adult-onset disorders. Having genomes captured as GS would ensure higher clinical sensitivity compared with ES and can readily be used for ancillary genetic information, such as pharmacogenetic and polygenic risk factors.

Mandatory NBS has improved early diagnoses of medically actionable conditions. However, current NBS approaches are limited to the RUSP, with 37 core conditions and 26 secondary conditions.⁶⁹ In the US, states offer limited, full, or enhanced screening for additional RUSP disorders depending on the program. Throughout the US, 0.3% of newborns (1 in 294) are found to be at risk for 31 NBS disorders,⁷⁰ with estimates reaching 0.8% (1 in 121) for 61 molecularly screened disorders in Pennsylvania (T. Donti, PhD, email communication, June 5, 2023). Newborn GS can expand early diagnostic capabilities beyond the disorders included in RUSP, but adaptability of GS to yield rapid results without causing additional distress or inconvenience to newborns or families is critical.^12,36,38 Our study demonstrated that DBS, as a specimen, can be successfully used for GS. Although our workflow was not developed for fast results, a strategy of integrating rapid GS testing with traditional NBS for early detection of 400 genetic conditions was recently published.³⁷ Finally, as evident from this consumer-driven testing, proactive newborn GS is receiving societal recognition. The GS option appears to be favored over ESGP by 2.75:1 based on our consumer preference estimation from the number of GS and ESGP orders received over the same 3-year period and weighted for the difference in the out-of-pocket costs. However, due to high out-of-pocket costs, accessibility is limited by socioeconomic class. Unless newborn GS is integrated population wide, access to GS is likely to further widen the diversity gap in health care. However, multifaceted barriers related to cost; health care system capacity; family education and support; and ethical, legal, and social implications would need to be addressed for implementing GS in NBS programs.^35,71,72,73 Incomplete penetrance and variable expressivity findings are very challenging to counsel and add significant burden on health care professionals and counseled families. Therefore, ongoing debate on what is reportable or actionable and what genes to include (informed by reports from proactive screening such as this) is key to finding the best approaches for future NBS.⁷⁴

Limitations

This study has some limitations. The main limitation is the absence of information from posttesting clinical follow-up, which is critical to assessing the clinical effects of these findings on the individuals and their families. Another limitation includes a possibility of ascertainment bias due to some families opting to screen their children on clinical suspicion, which was not obvious for the physician performing virtual pretest consultation based on medical records available and family narrative. Additionally, GS screening of healthy children involves many ethical, legal, and societal considerations that are not addressed in this report.

Conclusions

In this case series study, a significant proportion of apparently healthy children screened by GS were found to be at risk for a wide range of pediatric-onset conditions likely to be missed on limited gene panels. Many of these risks involve high-penetrance, often neurodevelopmental, disorders that may benefit from early interventions, leading to better prognosis and clinical outcomes.

Supplement 1.

eTable 1. Exome Sequencing Based Panel (ESGP) of 268 Genes Associated With Medically Actionable Pediatric Conditions

eTable 2. Pharmacogenomic Variants Included in Genome Sequencing (GS) Reports

eTable 3. Genes With Positive GS Findings and Their Associated Disease Categories

eTable 4. Specimen Type Distribution

eTable 5. Test and Sequencing Attributes

Supplement 2.

Data Sharing Statement