Statistical methods to disentangle genetic effects influencing infertility and early fetal viability with a genome-wide application

Genetic selection occurs at multiple stages before and during pregnancy. While parental genomes influence the probability of fertilization, the fetal genome, once established, plays a critical role in early fetal survival. However, when estimated separately, parental and fetal genetic effects may confound each other. To address this, we developed an extension of the case-parent triad design to jointly estimate the genetic contributions of the parents and the fetus. Our approach considers all offspring as carriers of the trait "fetal survival". As use of assisted reproductive technology (ART) usually reflects fertility issues, we performed separate analyses on non-ART and ART family units, hypothesizing that parental and fetal effects differ between these groups. In the Norwegian Mother, Father, and Child Cohort Study, we had access to genotypes for approximately 43,000 family triads and dyads, including 1,336 offspring conceived through ART. In the non-ART sample, we identified genome-wide significant fetal effects on fetal survival for SNPs within regions harboring genes relevant to infertility and fetal development, such as MDC1, MICB, HCP5, and NOTCH4. These effects remained significant after adjusting for parental interaction effects, confirming their origin as fetal effects. When we replicated the analysis in the ART sample, we observed partial overlap in fetal effects with those identified in the non-ART sample. Parental interaction effects were observed in both the non-ART and ART samples, but the specific genetic associations differed between the groups. Notably, several SNPs associated with parental interaction effects in the ART sample mapped to genes previously implicated in male infertility, including ACTB, FSCN1, and RNF216. Our findings have broad implications for understanding the genetic architecture of infertility and fetal development. To support the interpretation of our results, we provide detailed descriptions of the models, highlighting their strengths and limitations.