Gene-based analyses of the maternal genome implicate maternal effect genes as risk factors for conotruncal heart defects.

Congenital heart defects (CHDs) affect approximately 1% of newborns. Epidemiological studies have identified several genetically-mediated maternal phenotypes (e.g., pregestational diabetes, chronic hypertension) that are associated with the risk of CHDs in offspring. However, the role of the maternal genome in determining CHD risk has not been defined. We present findings from gene-level, genome-wide studies that link CHDs to maternal effect genes as well as to maternal genes related to hypertension and proteostasis. Maternal effect genes, which provide the mRNAs and proteins in the oocyte that guide early embryonic development before zygotic gene activation, have not previously been implicated in CHD risk. Our findings support a role for and suggest new pathways by which the maternal genome may contribute to the development of CHDs in offspring.

Introduction

Congenital heart defects (CHDs) are the most common group of birth defects, with a prevalence of approximately 1% in live births [1]. CHDs are also the leading cause of birth defect-related mortality [2] and account for the largest percentage of birth defect-associated hospitalizations and healthcare costs [3]. As for many birth defects, the risk of CHDs is associated with several genetically-mediated, maternal phenotypes, including folate status, obesity, pregestational diabetes, chronic hypertension, and preeclampsia [4, 5]. These associations suggest that the maternal genotype may contribute to the risk of birth defects in offspring, independent of the maternal alleles transmitted to the child. For example, maternal genes involved in folate transport and metabolism may influence the availability of folate to the embryo, which in turn influences the risk of folate-related birth defects.

While there has been some interest in assessing the relationship between birth defects and maternal genotypes (e.g., methylenetetrahydrofolate reductase or MTHFR genotypes) [6-10], studies of the maternal genotype have considered a relatively small number of maternal phenotypes and are limited by gaps in our understanding of the genetic contribution to these phenotypes. Further, studies focused on maternal phenotypes ignore maternal genes that might act through alternate mechanisms to influence the risk of birth defects. For example, studies in model systems indicate that mutations in maternal effect genes (MEGs), which provide the mRNAs and proteins in the oocyte that guide early embryonic development before activation of the embryonic genome, can result in birth defects in offspring [11-13]. While genome-wide association studies (GWAS) provide a comprehensive, agnostic approach for identifying disease associations, only a few GWAS have focused on the maternal genotype [14-17]. Consequently, there is much to be learned about the role of maternal genes in determining the risk of birth defects such as CHDs.

We have previously conducted a single nucleotide polymorphism (SNP)-based GWAS of maternal genetic effects for conotruncal heart defects (CTDs) [14], which affect the cardiac outflow tracts [18] and account for approximately one-third of all CHDs [19]. Although we identified several maternal SNPs with suggestive evidence of association (p ≤ 10−5) with CTDs, no association was genome-wide significant (p < 5 × 10−8). Compared to SNP-based GWAS, gene-based GWAS has the advantage of a less stringent threshold for statistical significance. Furthermore, gene-based analyses can include both common and rare variants [20] and, therefore, capture a greater proportion of the within gene variation than SNP-based analyses, which generally exclude variants with minor allele frequencies (MAFs) less than 5% [21]. Given these advantages, we have undertaken gene-based GWAS and meta-analyses using data from two large CTD datasets to identify maternal genes associated with the risk of CTDs in offspring.

Materials and methods

Study subjects

The Children’s Hospital of Philadelphia (CHOP)

Patients with CTDs and their parents were recruited through the Cardiac Center at CHOP (1992–2010), under a protocol approved by the Institutional Review Board for the Protection of Human Subjects at CHOP [14, 15]. Adult participants provided written consent for themselves and their participating minor children.

Patients with the following diagnoses were eligible to be a CTD case: tetralogy of Fallot, persistent truncus arteriosus, D-transposition of the great arteries, double outlet right ventricle, ventricular septal defects (conoventricular, posterior malalignment, and conoseptal hypoplasia types), aortic-pulmonary window, interrupted aortic arch, and isolated aortic arch anomalies. Medical records, imaging (e.g., echocardiography and cardiac magnetic resonance imaging), and operative reports were used to confirm cardiac diagnoses. Potential cases were tested for the 22q11.2 deletion syndrome using fluorescence in situ hybridization, multiplex ligation-dependent probe amplification, or both, and those with a deletion were excluded [22]. Potential cases were also excluded if they had a clinically diagnosed chromosome abnormality or single-gene mutation.

Pediatric Cardiac Genomics Consortium (PCGC)

Patients with CTDs and their parents were recruited as part of the PCGC Congenital Heart Defect GEnetic NEtwork Study (2010–2012) [23]. Recruitment took place at five main (including CHOP) and four satellite clinical sites. Informed consent was obtained under protocols approved by the Institutional Review Board for each study site. Adult participants provided written consent for themselves and their participating minor children.

Patients recruited through the PCGC included those with the same CTD diagnoses as listed above. Cardiac diagnoses were confirmed through review of medical records and electronic case reports, and potential CTD cases were excluded if they had a clinically diagnosed chromosomal or genetic disorder. Participants recruited at CHOP as part of the PCGC do not overlap with the CHOP participants described above.

Genetic methods

Blood samples were collected from cases, and blood or saliva samples were collected from parents of cases. When blood collection was scheduled in conjunction with a surgical procedure, the sample was collected before any blood transfusion. DNA extraction was performed using standard techniques. Genome-wide microarray genotyping was performed at the CHOP Center for Applied Genomics [14]. The CHOP samples were genotyped using the Illumina HumanOmni-2.5, Illumina HumanHap550 (v2 or v3), or 610 BeadChip platforms, and the PCGC samples were genotyped on the Illumina HumanOmni-1 or 2.5 platforms.

Imputation and Quality Control (QC) procedures

Standard QC procedures were performed for each dataset using Plink version 1.07 [24] and have been previously described [25]. Before imputation, the genotype data were checked for strand and coding errors. Case-parent trios were removed if more than 1% of genotyped SNPs had Mendelian errors. Suspected duplicate samples were identified using pairwise identity-by-descent estimation, and samples with pi-hat greater than 0.6 were removed. Samples with genotyping rates less than 95% were also removed. In addition, SNPs with MAF less than 1%, genotyping rates less than 90%, and all non-autosomal variants were excluded.

Due to differences in microarray genotyping platforms, the CHOP and PCGC case-parent trios data were imputed separately. After the pre-imputation exclusions, the CHOP data from different platforms (HumanOmni-2.5, HumanHap550K v2, 550K v3, and 610K) were combined, and the SNPs present across all platforms (N = 283,977 SNPs) were used for imputation. Similarly, the PCGC data from different platforms (HumanOmni-1 and HumanOmni-2.5) were combined, and the SNPs present on both platforms (N = 624,419 SNPs) were used for imputation.

For each dataset, haplotypes were pre-phased using SHAPEIT2 v2.727 [26], and imputation was performed using Impute2 v2.3.0 [27] with pre-phased haplotype data from the 1000 Genomes Project (version: Phase-I integrated v3 variants set) as the reference population. A genotype was imputed, only if the posterior probability value exceeded 0.9, the default calling threshold for Impute2. After imputation, we excluded SNPs with poor imputation quality (Impute2 information metric score less than 0.8), or genotyping rates less than 90%. Samples with genotyping rates less than 95% were removed. Because we were interested in assessing both common and rare SNPs, the post-imputation QC procedures did not include restrictions based on MAFs.

Statistical analysis

Genome-wide gene-based analyses

Maternal genetic effects were evaluated using a case-control approach in which mothers and fathers from the CTD trios were considered as cases and controls, respectively. Genes were defined by their transcription start-stop positions, including untranslated regions (hg19 reference assembly) plus 1kb upstream and downstream. Analyses were conducted separately for the CHOP and PCGC datasets, using the sequence kernel association test for the combined effect of common and rare variants (SKAT-C) [28]. In this approach, separate scores were calculated for the common (MAF ≥ 5%) and rare (MAF < 5%) variants in each gene, and p-values were based on the weighted sum of these scores. We used the SKAT-C default parameters for weighting common and rare SNPs and evaluated all autosomal genes with at least one common and one rare variant in our data.

To control for population stratification bias, only the parents of non-Hispanic Caucasian CTD cases (based on self- or parental-report) were included in the analyses. As race/ethnicity was based on the report rather than genetic data, we adjusted for the first genotypic principal component. Genotypic principal components analyses were conducted using Golden Helix SVS version 8.1 (Golden Helix, Inc., Bozeman, Montana, USA; www.goldenhelix.com), using the default parameter settings (additive genetic model, MAF-based allele classification, and each marker data normalized by its theoretical standard deviation under Hardy Weinberg Equilibrium). A meta-analysis of the gene-based results from the CHOP and PCGC datasets was performed using Fisher’s combination of probability method [29]. For each analysis, the genomic inflation factor (λ) was calculated, and quantile-quantile (Q-Q) plots were constructed to check for deviation of the observed distribution of the test statistic from the expected null distribution.

An association was considered genome-wide significant if the meta-analysis p-value was less than the Bonferroni-corrected p-value, based on the number of genes evaluated. Genes with meta-analysis p < 10−3 were considered to have suggestive evidence of association. For genes with at least suggestive evidence of association in the meta-analysis, we considered those for which the meta-analysis p-value was lower than the p-values in the contributing datasets (i.e., the evidence for an association was stronger in the combined data than in either of the individual datasets) as candidate maternal CTD-related genes. When several contiguous genes met these criteria, which may reflect linkage disequilibrium between variants in genes that are in close proximity rather than independent association signals, we reviewed gene functions [30] to identify the most likely candidate gene in the region.

Gene-set enrichment analyses

Enrichment analyses using MetaCoreTM (Thomson Reuters, Life Science Research; https://portal.genego.com/metacore)), were performed for genes with meta-analysis p < 0.01 to identify enriched gene ontology (GO) processes, diseases (represented by biological markers), pathway maps, and pathway processes. For these analyses, a false-discovery rate (FDR)-corrected p < 0.05 was considered statistically significant.

Post hoc analyses of maternal effect genes (MEGs)

The most significant association in our meta-analysis was with a gene that has been suggested to be a MEG [31]. Given this finding, we elected to conduct an a posteriori, MEG gene-set analysis. For this analysis, we considered a gene to be an established MEG if it was included in at least one of two comprehensive reviews of the MEG literature (Table A of S1 File) [32, 33]. Fisher’s exact test was used to compare the proportion of established MEGs among all genes with meta-analysis p-values below and above a specified p-value cut-point (i.e., p < 0.05 versus p ≥ 0.05). A Fisher’s exact p < 0.05 was considered statistically significant. In addition, we cross-referenced the list of established MEGs with our list of candidate maternal CTD-related genes.

Results

In both the CHOP and PCGC datasets, the most common diagnosis in the offspring was the tetralogy of Fallot (Table 1). After QC exclusions, the CHOP dataset included 423 mothers and 380 fathers, and the PCGC dataset included 216 mothers and 219 fathers.

Table 1

Summary of the conotruncal heart defect phenotypes in the offspring of study subjects.

Conotruncal Heart Defect Phenotype	CHOP		PCGC
	N = 483	%	N = 244	%
Tetralogy of Fallot	196	40.6	73	29.9
D-transposition of the great arteries	95	19.7	52	21.3
Ventricular septal defects	90	18.6	34	13.9
Double outlet right ventricle	53	11.0	37	15.2
Isolated aortic arch anomalies	22	4.6	7	2.9
Truncus arteriosus	15	3.1	7	2.9
Interrupted aortic arch	6	1.2	6	2.4
Other	6	1.2	28	11.5

Abbreviations: CHOP, The Children’s Hospital of Philadelphia; PCGC, The Pediatric Cardiac Genomics Consortium.

The number of variants and genes included in the CHOP and PCGC datasets are summarized in Table 2. The Q-Q plots (S1 and S2 Figs) and genomic inflation factors (Table 2) for the analyses of the individual datasets provided little evidence for systematic bias (Tables C and D of S1 File). No genome-wide significant associations (p ⪅ 2.3 × 10−6) were detected in either dataset.

Table 2

Summary of the genetic data used in the analyses of the CHOP and PCGC datasets.

	Dataset (# mothers/# fathers)
	CHOP (423/380)	PCGC (216/219)
Total variants	5,605,644	6,815,834
Rare variantsa	3,500,915	4,574,369
Number of genes	21,187	22,002
Genomic inflation factor (λ)	1.06	1.05

Abbreviations: CHOP, The Children’s Hospital of Philadelphia; PCGC, The Pediatric Cardiac Genomics Consortium.

a Variants with minor allele frequency < 0.05.

Fisher’s method was used to conduct a meta-analysis of the SKAT-C p-values from the 20,962 genes (Table D of S1 File) that were analyzed in both the CHOP and PCGC datasets. The genomic inflation factor (λ = 1.07) and Q-Q plot provided little evidence of a systematic deviation from the expected distribution (Fig 1). Although no gene achieved genome-wide significance in the meta-analysis (Bonferroni-corrected p < 2.4 × 10−6), the meta-analysis p-value for the germ cell-specific gene, GGN, was of borderline significance (p = 7.1 × 10−6). The meta-analysis also provided suggestive evidence of association for an additional 30 genes (Table 3).

Fig 1

A quantile-quantile plot.

A quantile-quantile plot of meta-analysis p-values obtained by combining SKAT-C test p-values from genome-wide analyses of the CHOP and PCGC datasets.

Table 3

Maternal genes with suggestive evidence of association (p < 10−3) with conotruncal heart defects in the meta-analysis.

Gene	CHR	CHOP		PCGC		Meta-analysisa
		# of variants	p-value	# of variants	p-value	p-value
GGN	19	17	6.30 × 10−4	18	7.23 × 10−4	7.10 × 10−6
SPRED3	19	46	5.11 × 10−3	44	2.71 × 10−4	2.01 × 10−5
VARS2	6	88	7.29 × 10−6	84	4.76 × 10−1	4.71 × 10−5
FER1L6-AS1	8	78	5.66 × 10−6	153	7.30 × 10−1	5.54 × 10−5
LOC101927269	7	18	2.80 × 10−1	21	1.71 × 10−5	6.34 × 10−5
LOC151475	2	25	1.49 × 10−5	31	4.08 × 10−1	7.91 × 10−5
SUMO1	2	84	7.48 × 10−1	74	1.12 × 10−5	1.06 × 10−4
PSMD8	19	40	8.51 × 10−4	39	1.06 × 10−2	1.14 × 10−4
SPINT4	20	43	4.38 × 10−1	41	2.08 × 10−5	1.15 × 10−4
SLAIN2	4	292	1.11 × 10−2	262	8.89 × 10−4	1.23 × 10−4
YIF1B	19	52	3.20 × 10−2	58	3.12 × 10−4	1.25 × 10−4
CATSPERG	19	162	4.72 × 10−3	186	2.29 × 10−3	1.34 × 10−4
FER1L6	8	785	4.51 × 10−5	897	2.84 × 10−1	1.57 × 10−4
LOC101928565	1	113	4.45 × 10−2	163	3.14 × 10−4	1.70 × 10−4
SFTA2	6	33	1.01 × 10−4	31	2.73 × 10−1	3.18 × 10−4
PTPRF	1	342	3.56 × 10−1	317	7.95 × 10−5	3.25 × 10−4
PLXND1	3	224	1.26 × 10−3	208	2.35 × 10−2	3.38 × 10−4
TBC1D29	17	19	2.57 × 10−2	12	1.16 × 10−3	3.40 × 10−4
KDM4A	1	161	3.21 × 10−1	120	1.05 × 10−4	3.81 × 10−4
TNK2	3	100	9.79 × 10−2	99	3.77 × 10−4	4.14 × 10−4
ZSWIM3	20	110	2.52 × 10−1	98	1.72 × 10−4	4.80 × 10−4
LOC100505978	12	8	1.32 × 10−1	10	4.15 × 10−4	5.92 × 10−4
MYDGF	19	71	1.21 × 10−4	80	4.52 × 10−1	5.92 × 10−4
FTH1	11	14	1.93 × 10−4	8	3.05 × 10−1	6.33 × 10−4
WFDC13	20	27	6.72 × 10−1	26	9.08 × 10−5	6.53 × 10−4
WFDC3	20	89	5.44 × 10−1	77	1.24 × 10−4	7.13 × 10−4
H1FOO	3	50	2.16 × 10−2	46	3.50 × 10−3	7.92 × 10−4
TBX20	7	102	2.66 × 10−4	124	3.10 × 10−1	8.58 × 10−4
ZNF622	5	40	1.96 × 10−4	45	4.24 × 10−1	8.62 × 10−4
HPS3	3	193	4.54 × 10−4	193	1.90 × 10−1	8.91 × 10−4
STARD7-AS1	2	37	3.53 × 10−1	35	2.74 × 10−4	9.92 × 10−4

Abbreviations: CHR, chromosome; CHOP, The Children’s Hospital of Philadelphia; PCGC, The Pediatric Cardiac Genomics Consortium.

a The meta-analysis included 20,992 genes.

Of the 31 genes with suggestive evidence for association, ten had meta-analysis p-values lower than the p-values in either individual dataset. These ten genes included one pseudogene (TBC1D29P), one RNA gene (LOC101928565), and eight protein-coding genes. The eight protein-coding genes include two contiguous genes located at 3q22.1 (H1FOO and PLXND1); SLAIN2 at 4p11; and five genes located in an approximately 100,000 base-pair region of 19q13.2 (YIF1B, CATSPERG, PSMD8, GGN, and SPRED3) (Table 4). Based on their known functions (Table 4), the eight protein-coding genes do not appear to be strong candidates for maternal genes that act via a maternal phenotype (e.g., obesity and diabetes). However, the 3q22.1 region includes a known MEG, H1FOO (meta-p = 7.9 × 10−4), and the 19q13.2 region includes a gene that has been proposed to be a MEG, GGN (meta-p = 7.1 ×10−6). Consequently, we propose H1FOO and GGN, as well as SLAIN2 (the single associated gene in the 4p11 region) as the top candidate maternal CTD-related genes identified by our meta-analysis.

Table 4

Maternal protein-coding genes with meta-analysis p-values suggestive of association (p < 10−3) and lower than the p-values from the analysis of individual datasets.

Chr.	Gene	Positiona	Descriptionb	Maternal Effect Gene	CHOP p-value	PCGC p-value	Meta-analysis p-value
19q13.2	YIF1B	38,793,200–38,807,606	Membrane trafficking protein		3.20 × 10−2	3.12 × 10−2	1.25 × 10−4
	CATSPERG	38,825,443–38,862,589	Sub-unit of the sperm calcium channel, CATSPER		4.72 × 10−3	2.29 × 10−3	1.34 × 10−4
	PSMD8	38,864,190–38,875,464	Involved in ATP-dependent degradation of ubiquinated proteins		8.51 × 10−4	1.06 × 10−2	1.14 × 10−4
	GGN	38,873,992–38,879,668	Germ cell specific gene	Suggested	6.30 × 10−4	7.23 × 10−4	7.10 × 10−6
	SPRED3	38,879,840–38,891,523	Negative regulation of MAP kinase signaling		5.11 × 10−3	2.71 × 10−4	2.01 × 10−5
3q22.1	H1FOO	129,261,057–129,271,310	Oocyte specific member of the H1 histone family	Established	2.16 × 10−2	3.40 × 10−3	7.92 × 10−4
	PLXND1	129,273,056–129,326,582	Cell surface receptor for semaphorins		1.26 × 10−3	2.35 × 10−2	3.38 × 10−4
4p11	SLAIN2	48,342,613–48,429,215	Promotes cytoplasmic microtubule nucleation and elongation		1.11 × 10−2	8.89 × 10−4	1.23 × 10−4

Abbreviations: CHOP, The Children’s Hospital of Philadelphia; PCGC, The Pediatric Cardiac Genomics Consortium.

a Gene transcription start/stop positions (hg19) plus 1 kb upstream and downstream.

b Gene descriptions obtained from GeneCards: https://www.genecards.org/.

Gene-set enrichment analyses

In enrichment analyses of genes with meta-analysis p < 0.01 (N = 204 genes), no pathway map or pathway process was significant. However, there was evidence of enrichment (FDR p < 0.05) for 17 GO processes (Table E of S1 File), including several related to transmembrane transport in general, and calcium ion transport in particular (e.g., GO:1903169, regulation of calcium ion transmembrane transport, p = 2.9 × 10−2). There was also evidence for enrichment of genes for biological markers associated with 24 disease processes including, diseases of proteostasis (e.g., proteostasis deficiencies, p = 8.2 × 10−3) and hypertension (FDR p = 3.0 × 10−2) (Table F of S1 File).

Post hoc analyses of MEGs

Given that the most significant association in our meta-analysis was with GGN (p = 7.1 x 10−6), a gene that has been suggested to be a MEG [31], we elected to conduct an a posteriori, MEG gene-set analysis. Based on two comprehensive reviews of the MEG literature [32, 33], we identified a list of 60 MEGs (Table A of S1 File). In our meta-analysis six of the genes on this list had p < 0.05 (HF1OO, p = 0.0008; KMT2D, p = 0.015; TP73, p = 0.026; BNC1, p = 0.034; ZAR1, p = 0.036; and RNF2, p = 0.0497). Although GGN also had a meta-analysis p < 0.05, this gene was not included in either of the review articles and was therefore omitted from these analyses. The identification of six MEGs with meta-analysis p < 0.05 represents a 2.3-fold enrichment, which is of borderline significance (Fisher’s exact p = 0.057) based on the standard p-value cut-off for a single statistical test (i.e. p < 0.05).

Discussion

Our genome-wide, gene-based analyses of common and rare variants provide suggestive evidence that maternal genes are associated with the risk of CTDs in their offspring. Based on the analyses of individual genes, we identified three candidate CTD-related maternal genes, H1FOO, GGN, and SLAIN2, and propose that these genes are most likely to influence CTD-risk via effects on early embryonic development.

H1FOO (meta-p = 7.1 × 10−6), a known MEG [33], is an oocyte-specific member of the linker histone H1 family [34]. Genes in this family are involved in the determination of higher-order chromatin structure and gene transcription. Knockdown studies of H1foo in mouse one-cell embryos indicate that maternal H1foo influences the progression of DNA replication by reducing the deposition of H3 in the perinuclear region of the male pronucleus, and significantly delays the timing of cleavage into a two-cell embryo [35].

The suspected MEG, GGN, is thought to be involved in DNA repair and is characterized as a germ cell-specific gene. GGN is expressed at high levels in the adult testis [36], and at lower levels in the adult ovary and somatic tissues, as well as in Metaphase-II (MII) oocytes and early embryos [36, 37]. Evidence that GGN may function as a MEG is based on the timing of the loss of viability in mouse Ggn embryos. Specifically, Ggn embryos are present in expected numbers at the two-cell stage but are rarely observed at the morula stage and absent by embryonic day 7.5, consistent with the loss of viability following the depletion of maternal Ggn mRNA stores [31].

Although SLAIN2 has not previously been implicated as a MEG, SLAIN2 mRNA is abundant in both MII oocytes and one-cell embryos and declines thereafter [37]. Further, SLAIN2 is involved in microtubule dynamics and organization [38], which are essential for several post-fertilization processes, including meiotic spindle assembly, separation of the parental genomes, and pronuclei migration [39-42]. Hence, both the expression pattern and known functions of SLAIN2 are compatible with a potential role as a MEG.

Additional evidence that MEGs may be associated with CTDs in offspring is provided by the observed 2.3-fold enrichment of established MEGs among genes with p < 0.05 in our meta-analysis. The established MEGs with meta-analysis p < 0.05 include: H1FOO (discussed above); the transcriptional regulators, BNC1 and KMT2D; RNF2, which is involved in chromatin remodeling; and, TP73 and ZAR1, which are involved in cell cycle regulation [32]. Although MEGs have not previously been implicated as potential maternal risk factors for CTDs, studies in model systems demonstrate that mutations in MEGs can have a range of consequences for offspring, including embryonic lethality, developmental delay, and congenital malformations [11-13]. Similarly, women carrying a MEG mutation (e.g., NLRP5, NLRP7, and PADI6) experience a range of reproductive outcomes, including hydatidiform moles, periods of infertility, reproductive loss, offspring with multi-locus imprinting disorders, and unaffected children [43-46]. Although somewhat anecdotal, it is of interest that one (of five) woman with an NLRP5 mutation, ascertained following the birth of a child with a multi-locus imprinting disorder, also had a child with an isolated (i.e., apparently non-syndromic) CHD (atrial septal and ventricular defects) [43].

The observed enrichment of genes mapping to GO processes related to ion transmembrane transport, and specifically to calcium ion transport, could also be driven by MEGs. Although a detailed understanding of the genetic regulation of these oscillations is lacking, the known MEG, NLRP5 (also known as MATER), is required for calcium homeostasis. Specifically, oocytes from mouse Mater hypomorphs exhibit lower first peak amplitudes and higher frequencies of calcium oscillations (as compared to wild-type oocytes), likely due to a reduction in calcium stores in the endoplasmic reticulum [47].

Our analyses also identified the enrichment of genes related to hypertension. Maternal pregestational hypertension is associated with an increased risk of several birth defects, including, but not limited, to CHDs [48-51]. These associations appear to be independent of medications taken for the treatment of hypertension [48, 49], suggesting either that maternal hypertension, per se, has a negative impact on development (e.g., via an effect on blood flow to the uterus) or that hypertension and birth defects share common risk factors (e.g., genes with pleiotropic effects).

Our analyses also identified enrichment of genes related to proteostasis deficiency and diseases associated with protein misfolding and aggregation (e.g., amyotrophic lateral sclerosis). During pregnancy, the accumulation of misfolded proteins in body fluids and the placenta is associated with preeclampsia, a maternal condition that is also associated with an increased risk of birth defects, including CHDs [50-54].

The results of our study must be viewed in light of both its strengths and limitations. We used a case-control study design, in which we compared the mothers (cases) and fathers (controls) of individuals with CTDs, to identify maternal CTD-related genes. Compared to a case-control design using unrelated, female controls, our design has the advantage of controlling for the potentially confounding effects of the genotype inherited by the child but is subject to bias arising from differences in allele frequencies between males and females. However, sex differences in allele frequencies appear to be uncommon (< 1% of variants) in autosomal genes [55]. In addition, while our analyses assess whether there are gene-level differences between mothers and fathers, they do not indicate which group might carry more (or less) disease-related alleles. Our observed associations could, therefore, be driven by paternal rather than maternal effects. However, since embryonic development prior to zygotic gene activation is primarily driven by maternal gene products, and the maternal genome has a direct effect on the in utero environment, any true associations detected in our study are most likely due to maternal genes. Nonetheless, additional studies (e.g., in model systems) will be required to confirm and establish the mechanisms underlying these observed associations.

Our analyses were based on two large CTD datasets that were ascertained in the United States using similar recruitment, and systematic case confirmation (phenotyping) approaches. Furthermore, our gene-based analyses had a lower multiple-testing burden than SNP-based GWAS. However, our sample sizes were relatively small for our genome-wide approach, and the criterion for achieving statistical significance (corrected-p ~ 2.5 × 10−6) remained quite high. Consequently, associations with maternal CTD-related genes may have been missed in our analyses due to low power. Genes with suggestive evidence of association and genes associated with enriched terms, therefore, appear to be strong targets for further investigations of the maternal genetic contribution to CTDs.

In our analyses, we combined data across different CTD phenotypes, which could have obscured associations if the maternal contribution to individual phenotypes is distinct. However, the predominance of evidence suggests that maternally mediated risk factors tend to be related to a broad spectrum of malformations. For example, maternal hypertension, preeclampsia, diabetes, and obesity are all associated with a spectrum of cardiac and non-cardiac malformations. Hence, for studies of maternal risk factors, the potential for improved power resulting from analyses of similar birth defects (e.g., the various CTD phenotypes) outweighs concerns regarding the potential impact of phenotypic heterogeneity.

This study is the first gene-based GWAS of maternal genotypes and CTDs. We have, however, previously conducted SNP-based, common-variant GWAS and meta-analysis using the same datasets as in the current gene-based analyses [14]. In our SNP-based meta-analysis, we identified several variants with suggestive evidence of association (p ≤ 10−5); however, none were located in, or within 1kb of the genes with suggestive evidence of association in the current, gene-based analyses. Based on these same two datasets, we have also reported that the risk of CTDs is associated with a maternal genetic risk score for hypertension [10]. However, only one variant included in the genetic risk score falls within a gene that was included in our enrichment analyses (i.e., rs11862778 in MTHFR). Hence, these two analyses appear to provide largely independent evidence that genes related to maternal hypertension are associated with the risk of CTDs.

In conclusion, our analyses provide provocative new insights into the potential influence of the maternal genome on embryonic development. While our results are specific to CTDs, both maternal conditions (e.g., hypertension) and MEGs are associated with a range of adverse reproductive outcomes, suggesting that our findings may have much broader implications for the understanding of birth-defect etiology. Further, our findings suggest a link between birth defects and other adverse pregnancy outcomes (e.g., reproductive loss and infertility). Confirmation of such a link would have broad implications for reproductive counseling and planning. Given these initial, compelling findings, additional studies of the relationship between the maternal genome and birth defects are warranted.

Quantile-quantile plot of SKAT-C test gene-based p-values in the CHOP cohort (genomic inflation factor = 1.06).

(PDF)

Click here for additional data file.

Quantile-quantile plot of SKAT-C test gene-based p-values in the PCGC cohort (genomic inflation factor = 1.05).

(PDF)

Click here for additional data file.

Table A. Human homologs of mammalian maternal effect genes identified in reviews by Condic (2016) and Zhang and Smith (2015). Table B. A gene-based analysis of the CHOP cohort using the SKAT-C test. The mothers and fathers of patients with CTDs were considered as 'cases' and 'controls,' respectively, for this analysis. Table C. A gene-based analysis of the PCGC cohort using the SKAT-C test. The mothers and fathers of patients with CTDs were considered as 'cases' and 'controls,' respectively, for this analysis. Table D. A gene-based meta-analysis of SKAT-C test results from the CHOP and PCGC cohort analyses. Table E. Gene Ontology (GO) processes from MetaCore enrichment analysis of top genes (meta-analysis p < 0.01) from the SKAT-C test, and color-coded GO process clusters identified through REVIGO. Table F. Enriched diseases (by biological markers) in the MetaCore enrichment analysis of top genes (meta-analysis p < 0.01) from the SKAT-C test.

(XLSX)

Click here for additional data file.

Transfer Alert

This paper was transferred from another journal. As a result, its full editorial history (including decision letters, peer reviews and author responses) may not be present.

7 Apr 2020

PONE-D-20-05538

Gene-based analyses of the maternal genome implicate maternal effect genes as risk factors for conotruncal heart defects

PLOS ONE

Dear Dr. Mitchell,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by May 22 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

David Scott Winlaw, MBBS MD FRACS

Academic Editor

PLOS ONE

Additional Editor Comments (if provided):

Thank you for submitting this work. The reviewers have raised a number of important concerns and limitations of the study that require addressing and modification of the manuscript, mostly allowing for readers to better understand the context and outcome of the study. Greater clarity in the introduction of why you have have chosen this study methodology, and a broader contextualization of the results for a scientific and medical audience are required before further consideration of publication. There seems to be a gap between the stated aims of the paper, and the validity of the argument in supporting the hypotheses. There is an over-reliance on statistical significance supporting relevance rather than a cohesive argument supporting the role of particular genes in the development of congenital heart disease.

Revision of the manuscript taking into consideration the reviewer comments may strengthen the paper, a point by point rebuttal and outline of changes made will allow further consideration of the paper.

Journal Requirements:

When submitting your revision, we need you to address these additional requirements:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. One of the noted authors is a group or consortium "Pediatric Cardiac Genomics Consortium". In addition to naming the author group, please list the individual authors and affiliations within this group in the acknowledgments section of your manuscript. Please also indicate clearly a lead author for this group along with a contact email address.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This study begins to address the interesting and important question of maternal effect gene influence on congenital heart disease in the child. The is the first report of a burden test looking for enrichment in maternal variants over this in the father. The manuscript is well-written and includes a detailed methods section. It also reports some interesting possible associations that could be verified and followed up by subsequent studies. The discussion section is valid on the whole but it is long. Statistical analysis appear to be performed to a high standard. All data presented in this manuscript have been made available and the appropriate ethics appear to have been obtained.

Pathway analysis (Gene Set Enrichment Analysis (GSEA)) has been mentioned (in methods and in the context of meta-analysis) but not it is not clear why GSEA was not performed on individual datasets (as was done for gene-based burden testing). What relevant gene sets were tested?

More importantly, if the aim of this work was to demonstrate a significant enrichment in variants in MEGs in maternal versus paternal genomes, then a MEG list could have tested alone by GSEA, removing the need for multiple comparison correction, and increasing the chance of a significant result.

Did the authors consider filtering those variants included in the burden test based on receiving a high score by pathogenicity predictors such as CADD and polyphen-2, as has been reported in the literature?

In the meta-analysis, the authors report a 2.3-fold enrichment of known MEGs. Could the authors clarify (in the text) if this is significant? It would appear not if the fisher exact figure of p=0.057 pertains to this fold increase. The first sentence of the discussion is not really correct if this is the case.

The term gene-based GWAS is confusing. Use of a more established terms such as rare variant association testing or burden testing could be used and would make it clearer what type of analysis the was being performed. If the authors still prefer the term “gene-based GWAS” then other terms by which it is known (above) could also be provided in brackets. The term Gene Set Enrichment Analysis (GSEA) could also used in places where pathways and gene lists are analysed.

It is somewhat surprising that the placenta is not mentioned in this manuscript. It would seem likely that some MEGs important for CHD include genes relevant to placenta formation, given that disruption of genes in mice that cause placental defects tend to also have heart and vascular defects (Perez-Garcia et al Nature 2018). Comment on the lack enrichment in placental genes could be made in the discussion. Is there a list of genes important for maternal placenta formation that could be used in GSEA?

The author’s rightly point out that true associations are likely maternal rather than paternal associations, however the manuscript does not present those genes enriched for variants in the paternal genomes, which would be useful as a means to judge the number of false associations observed in the analysis.

Reviewer #2: Just a few comments/queries to hopefully improve clarity and transparency:

Page 9 line 181: Replace "an established MEG" with "a previously established MEG"

Page 11 line 218-220 and page 13 lines 261-266: I did not understand why you are including SLAIN2 as one of your "candidate maternal CTD-related genes"? What is it about SLAIN2 (as opposed to the 5 genes on 19q13.2) that puts it into this category?

Page 12 line 239: Replace "provide evidence" with "provide suggestive evidence" or "provide preliminary evidence" or something similar. (As the actual level of evidence seems pretty weak).

Page 13 line 261: Replace "not been implicated" with "not been previously implicated"

Page 16 lines 335-336. It would have been interesting to see results for specific phenotypes (particularly the ones with the largest sample sizes, namely the top 4 phenotypes listed in Table 1). Have you considered doing this? I would not insist on it for this publication, but worth considering in the future...

Page 34 Table 2: Please include the actual p values for the genes with p<10-3 listed (as separate columns, or in brackets after the gene name), as was done in Table 3.

Reviewer #3: Not novel, the authors have recently published a 2019 paper examining the same data set for GWAS/Meta-analyses of CHD in PLOS One (https://doi.org/10.1371/journal.pone.0219926) Not cited in current paper (CHOP trios and PGCG Trios)

While this current manuscript focuses on maternal effect genes, these genes did not reach genome wide significance in the previous paper although the methods highly similar.

It’s worth noting significant portions of the papers overlap with the authors precious publications

In the introduction, the authors state that they have previously conducted SNP based GWAS on this dataset but did not identify any loci with genome wide significance, and thus opted to conducted gene based GWAS as it allowed them to loosen the stringent thresholds for statistical significance while including both common and rare variants.

This coupled with the random urge to look at maternal contributions to CHD (these authors have previously examined maternal contributions in a 2014 PLos One paper and found no significant associations initial 2014 CHOP Trios study, they studied maternal contributions in a 2017 using CHOP trios and LVOTD Trios) seems to be “fishing” especially in light of the 2019 paper

The current paper under review used similar methodology, but looked instead at maternal affect genes, using fathers of the patients as controls. the use of the fathers as controls, completely ignores their contribution to the overarching CHD phenotype making them a less that optimal control especially in light of previous analyses with the proper control samples (2019 paper, 2017 paper and original 2014 paper)

Finally, the authors failed to cite a 2013 CHD GWAS by Cordell et al in Nature Genetics that was very well powered (1995 cases, 5159 controls) that examined 3 major CHD categories together then separately. The authors of this study concluded “Our work, therefore, adds to recent data from studies of CNVs, suggesting that genetic associations with CHD have a considerable degree of phenotypic specificity1”

� It appears that the authors of this paper under review have failed to recognize the heterogeneity within the CHD phenotype although in several of the previous papers, they do seem to be aware of it. A 2015 paper by the same group which examined left cardiac malformations via GWAS did identify several associated loci with genome wide significance.

� It is not clear why the authors insist on examining the CHD as a single phenotype when in the 2018 cohort description for the PCGC trios, they state that CHD is a “broad spectrum of malformations.”

� Further they claim in the manuscript under review the interpretations of their data are hampered by: limited understanding of the mechanisms underlie associations between maternal conditions and birth defects (hello placenta how are you doing? Not well my dear) and lack of detailed annotations specific to the role of maternal genes in offspring development “little is known about mammalian maternally expressed genes” (genes expressed in the egg prior to zygotic gene activation) a google search produced a number of manuscripts on this topic of note is a well cited review by Zhang and Smith 2016, Maternal Control of early embryogenesis in mammals – a review Table 1 lists many genes along with their citations.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

19 May 2020

Editor Comments

The reviewers have raised a number of important concerns and limitations of the study that require addressing and modification of the manuscript, mostly allowing for readers to better understand…

See below for our responses to the reviewer comments. We believe that the revisions we have made to the manuscript address the broader concerns expressed by the Editor.

When submitting your revision, we need you to address these additional requirements:

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming.

The revised manuscript has been edited to meet PLOS ONE’s style and file naming requirements.

Tables and Figures have been relocated so that they follow the paragraph in which they are first cited. We did not track these changes, as doing so would likely make it more difficult to follow the more substantive changes.

We note that you have indicated that data from this study are available upon request.

The data used in the analyses described in this manuscript were derived from three studies. These studies have all been registered through dbGap and data from two have been uploaded:

Pediatric Cardiac Genomics Consortium:

https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001194.v2.p2

CHOP pediatric controls:

https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000490.v1.p1

Data from the third study are being prepared for submission to the related dbGap project.

CHOP CTD trios:

https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000881.v1.p1

This information has been added to the Data Availability section of the manuscript.

One of the noted authors is a group or consortium "Pediatric Cardiac Genomics Consortium". In addition to naming the author group, please list the individual authors and affiliations within this group in the acknowledgments section of your manuscript. Please also indicate clearly a lead author for this group along with a contact email address.

We have moved the individual authors (names and affiliations) from the Pediatric Cardiac Genomics Consortium from Table A of Supplemental Files to the acknowledgements section of the manuscript. A contact author has also been designated and her email has been provided.

Note: Line references provided below refer to line numbers in the marked-up version of the revised manuscript.

Reviewer #1:

The discussion section is valid on the whole but it is long.

We have reduced the level of detail provided in several sections of the Discussion.

Pathway analysis (Gene Set Enrichment Analysis (GSEA)) has been mentioned (in methods and in the context of meta-analysis) but not it is not clear why GSEA was not performed on individual datasets (as was done for gene-based burden testing).

For all of our analyses, we have emphasized the results from the meta-analysis, because they are (by definition) based on larger numbers and, therefore, provide improved powered to detect associations relative to the individual datasets. However, based on this comment, we realized that our inclusion of the section titled ‘Gene-based GWAS of individual datasets,’ detracted from this emphasis. Consequently, we have removed this section (and modified the related Table 2) from the Results section of the manuscript. We have, however, retained the results of the analyses of the individual studies in the supplemental materials.

What relevant gene sets were tested?

Gene-set enrichment analyses were performed using MetaCore and included evaluation of MetaCore’s manually-curated gene ontology processes, diseases, pathway maps, and pathway processes (see lines 189-194. The analyses conducted using MetaCore were agnostic (i.e., we did not pre-specify processes of interest).

More importantly, if the aim of this work was to demonstrate a significant enrichment in variants in MEGs in maternal versus paternal genomes, then a MEG list could have tested alone by GSEA, removing the need for multiple comparison correction, and increasing the chance of a significant result.

Our MEG gene-set analysis was not planned a priori, but rather was motivated by the gene-level results in the meta-analysis, where the top gene, GGN, was a proposed MEG. To clarify this point, we have moved the description of the MEG gene-set analysis to a new section in Results, titled ‘Post hoc analyses of maternal effect genes (MEGs)’.

Did the authors consider filtering those variants included in the burden test based on receiving a high score by pathogenicity predictors such as CADD and polyphen-2, as has been reported in the literature?

Our analyses were based on comprehensive coverage of both common and rare variants in each gene. We did not consider or undertake analyses based on CADD or similar scores, as such an approach would dramatically decrease the coverage of each gene.

In the meta-analysis, the authors report a 2.3-fold enrichment of known MEGs. Could the authors clarify (in the text) if this is significant? It would appear not if the fisher exact figure of p=0.057 pertains to this fold increase. The first sentence of the discussion is not really correct if this is the case.

As noted above, this analysis is now clearly identified as being a post hoc analysis motivated by the results of our gene-level analysis. Further, in the Results section, we now state that the observed 2.3-fold enrichments “is of borderline significance (Fisher’s exact p = 0.057)” (lines 322-324). Finally, based on a revision suggested by Reviewer 2, the first sentence of the Discussion section has been modified to read “Our genome-wide, gene-based analyses of common and rare variants provide suggestive evidence…” (modified text in italics).

The term gene-based GWAS is confusing. Use of a more established terms such as rare variant association testing or burden testing could be used and would make it clearer what type of analysis the was being performed.

Prior to selecting a title for this manuscript, we reviewed related literature and found several publications that used the “gene-based GWAS” terminology. We believe it is important to retain the reference to GWAS, to clarify the scope of our analyses and to include the qualifier, “gene-based”, to distinguish our analyses from SNP-based GWAS. Consequently, we have elected to retain the “gene-based GWAS” terminology.

If the authors still prefer the term “gene-based GWAS” then other terms by which it is known (above) could also be provided in brackets. The term Gene Set Enrichment Analysis (GSEA) could also used in places where pathways and gene lists are analyzed.

In response to this comment, sections in both Methods and Results that were labeled as ‘Enrichment and gene-set analyses’ have been re-labeled as ‘Gene set enrichment analyses.’

It is somewhat surprising that the placenta is not mentioned in this manuscript. It would seem likely that some MEGs important for CHD include genes relevant to placenta formation, given that disruption of genes in mice that cause placental defects tend to also have heart and vascular defects (Perez-Garcia et al Nature 2018). Comment on the lack enrichment in placental genes could be made in the discussion. Is there a list of genes important for maternal placenta formation that could be used in GSEA?

As noted above, our initial analyses were focused on an agnostic, comprehensive assessment of genes and gene sets. Based on the results of the agnostic gene-level analyses, in which the top gene (GGN) is a suspected MEG, MEGs were assessed in post hoc analyses. As the gene-level analysis did not provide obvious links to the placenta, no post hoc analyses of placental genes were undertaken.

The author’s rightly point out that true associations are likely maternal rather than paternal associations, however the manuscript does not present those genes enriched for variants in the paternal genomes, which would be useful as a means to judge the number of false associations observed in the analysis.

The gene-level analyses identify whether a gene is associated with the outcome but do not provide an estimate of the direction or magnitude of the association. Thus, in this study, the analyses indicate whether (or not) there is a difference between mothers and fathers of cases, but do not indicate which group might carry more (or less) disease-related alleles. We have clarified this point in the Discussion section (text starting on line 409).

Reviewer #2:

Page 9 line 181: Replace "an established MEG" with "a previously established MEG"

Given the context within which this statement occurs, we believe the addition of “previously” is not required. Specifically, in the same sentence, we define that an established MEG is one that is included in at least one of two published review articles.

Page 11 line 218-220 and page 13 lines 261-266: I did not understand why you are including SLAIN2 as one of your "candidate maternal CTD-related genes"? What is it about SLAIN2 (as opposed to the 5 genes on 19q13.2) that puts it into this category?

In our Statistical Methods section we state that “For genes with at least suggestive evidence of association in the meta-analysis, we consider whether the meta-analysis p-value for the gene was lower than the p-values in the contributing datasets (i.e., the evidence for association was stronger in the combined data than in either of the individual datasets)…” Consequently, SLAIN2 is included as a proposed maternal CTD-related gene because it is the only gene on chromosome 4 with suggestive evidence of association in the meta-analysis, and a meta-analysis p-value that is lower than the p-values obtained in either of the individual studies.

Because the five genes on chromosome 19 (as well as the two genes on chromosome 3) that met these criteria are contiguous, it is likely that the observed associations are not independent, but rather reflect linkage disequilibrium between variants in different genes. Consequently, for each of these two regions, we elected the gene for which we believe there is the strongest evidence that it might act through the maternal genotype. We have revised the Method section (text starting on line 177) to clarify our rationale for the selection of our candidate maternal CTD-related genes in these two regions.

Page 12 line 239: Replace "provide evidence" with "provide suggestive evidence" or "provide preliminary evidence" or something similar. (As the actual level of evidence seems pretty weak).

We have revised this text by adding the word “suggestive”.

Page 13 line 261: Replace "not been implicated" with "not been previously implicated"

We have made this suggested revision.

Page 16 lines 335-336. It would have been interesting to see results for specific phenotypes (particularly the ones with the largest sample sizes, namely the top 4 phenotypes listed in Table 1). Have you considered doing this? I would not insist on it for this publication, but worth considering in the future...

We agree that analyses of individual CTD phenotypes would be of interest. However, particularly when assessing maternal genotypes, we believe that there is strong rationale for analyzing CTDs as a group (see Discussion, paragraph starting on line 447). For these reasons, and concerns about small sample sizes, we have not conducted analyses of the individual phenotypes.

Page 34 Table 2: Please include the actual p values for the genes with p<10-3 listed (as separate columns, or in brackets after the gene name), as was done in Table 3.

In response to a comment from Review 1, we have opted to focus on the findings from the meta-analysis. Consequently, Table 2 has been modified and no longer includes the list of suggestive genes. Exact p-values for all genes assessed in the individual studies have, however, been retained in the supplemental materials.

Reviewer #3:

Not novel, the authors have recently published a 2019 paper examining the same data set for GWAS/Meta-analyses of CHD in PLOS One

The referenced paper is based on the same datasets, but describes analyses of the association between CTDs and the inherited genotype. Consequently, the analyses described in that paper are not directly relevant to the analyses presented in this paper. We have, however, provided a reference for our prior SNP-level GWAS of the maternal genotype and CTDs (Agopian et al 2017).

While this current manuscript focuses on maternal effect genes, these genes did not reach genome wide significance in the previous paper although the methods highly similar.

True, but we would not expect that maternal genes associated with CTDs would be the same genes that are associated with CTDs via the inherited genotype.

It’s worth noting significant portions of the papers overlap with the authors precious publications

The analyses described in this paper are based on the same datasets that have been used in prior publications. Thus, there is overlap in the description of the study subjects and data collection and processing procedures. As we have not previously published on gene-level analyses of the maternal genotype, the results presented in this manuscript do not overlap with any of our prior work.

In the introduction, the authors state that they have previously conducted SNP based GWAS on this dataset but did not identify any loci with genome wide significance, and thus opted to conducted gene based GWAS as it allowed them to loosen the stringent thresholds for statistical significance while including both common and rare variants.This coupled with the random urge to look at maternal contributions to CHD (these authors have previously examined maternal contributions in a 2014 PLos One paper and found no significant associations initial 2014 CHOP Trios study, they studied maternal contributions in a 2017 using CHOP trios and LVOTD Trios) seems to be “fishing” especially in light of the 2019 paper.

In the introduction, we state that GWAS “provide a comprehensive, agnostic approach”. While this approach may be viewed as “fishing”, GWAS have had a major, positive impact on our understanding of the genetic contribution to complex traits.

The suggestion that our evaluation of the maternal contribution to CTD risk represents a “random urge” is incorrect. Evaluation of the maternal genotype is a topic of some interest within the birth defects research community and has been a major research focus for our group for over two decades. Our work in this domain has been and continues to be funded by the NIH, and we have multiple publications describing both our methodologic and applied work on this topic.

The current paper under review used similar methodology, but looked instead at maternal affect genes, using fathers of the patients as controls. the use of the fathers as controls, completely ignores their contribution to the overarching CHD phenotype making them a less that optimal control especially in light of previous analyses with the proper control samples (2019 paper, 2017 paper and original 2014 paper)

We assume that the reviewer is referring to our prior published studies. As noted above, some of these publications (e.g., Sewda et al. 2019) are based on studies of the inherited genotype and are not relevant to this submission. Our prior SNP-based studies of the maternal genotype (2014 and 2017) used a trio-based approach in which the maternal genotype is assessed by comparison of reciprocal mating types (e.g., mother AA x father aa versus mother aa versus father AA) – so these comparisons were also based on data from mothers and fathers.

We have expanded our Discussion of the strengths and limitations of our analytic approach in response to the last comment provided by Reviewer 1 (see response above and paragraph starting on line 402). In this section, we indicate that the comparison of mothers to fathers, in a case-control framework, appropriately controls for the correlation between the parental and inherited genotypes. Specifically, because the parent-child genetic correlation is the same for mothers and for fathers, genes that are associated with CTDs solely through the inherited genotype would not be identified using this design.

Finally, the authors failed to cite a 2013 CHD GWAS by Cordell et al in Nature Genetics that was very well powered (1995 cases, 5159 controls) that examined 3 major CHD categories together then separately. The authors of this study concluded “Our work, therefore, adds to recent data from studies of CNVs, suggesting that genetic associations with CHD have a considerable degree of phenotypic specificity1”

The Cordell paper provides an evaluation of the inherited genotype and, therefore, is not directly relevant to this study.

It appears that the authors of this paper under review have failed to recognize the heterogeneity within the CHD phenotype although in several of the previous papers, they do seem to be aware of it. A 2015 paper by the same group which examined left cardiac malformations via GWAS did identify several associated loci with genome wide significance.

It is not clear why the authors insist on examining the CHD as a single phenotype when in the 2018 cohort description for the PCGC trios, they state that CHD is a “broad spectrum of malformations.”

See our response to the second to last comment from Reviewer 2.

Further they claim in the manuscript under review the interpretations of their data are hampered by: limited understanding of the mechanisms underlie associations between maternal conditions and birth defects (hello placenta how are you doing? Not well my dear) and lack of detailed annotations specific to the role of maternal genes in offspring development “little is known about mammalian maternally expressed genes” (genes expressed in the egg prior to zygotic gene activation) a google search produced a number of manuscripts on this topic of note is a well cited review by Zhang and Smith 2016, Maternal Control of early embryogenesis in mammals – a review Table 1 lists many genes along with their citations.

We have removed the referenced text from the revised manuscript and have moved the remainder of the text from that paragraph to an earlier section of the discussion (lines 365-375).

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step

Figure 1 has been evaluated for PLOS requirements using PACE. The remaining figures appear in the Supporting Information files.

26 May 2020

Gene-based analyses of the maternal genome implicate maternal effect genes as risk factors for conotruncal heart defects

PONE-D-20-05538R1

Dear Dr. Mitchell,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

David Scott Winlaw, MBBS MD FRACS

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for revising your manuscript, this will make an excellent contribution.

Reviewers' comments:

29 May 2020

PONE-D-20-05538R1

Gene-based analyses of the maternal genome implicate maternal effect genes as risk factors for conotruncal heart defects

Dear Dr. Mitchell:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor David Scott Winlaw

Academic Editor

PLOS ONE