Purpose: The application of genomic sequencing to investigate unexplained death during early human development, a form of lethality likely enriched for severe Mendelian disorders, has been limited. Methods: In this study, we employed exome sequencing as a molecular autopsy tool in a cohort of 44 families with at least one death or lethal fetal malformation at any stage of in utero development. Where no DNA was available from the fetus, we performed molecular autopsy by proxy, i.e., through parental testing. Results: Pathogenic or likely pathogenic variants were identified in 22 families (50%), and variants of unknown significance were identified in further 15 families (34%). These variants were in genes known to cause embryonic or perinatal lethality (ALPL, GUSB, SLC17A5, MRPS16, THSD1, PIEZO1, and CTSA), genes known to cause Mendelian phenotypes that do not typically include embryonic lethality (INVS, FKTN, MYBPC3, COL11A2, KRIT1, ASCC1, NEB, LZTR1, TTC21B, AGT, KLHL41, GFPT1, and WDR81) and genes with no established links to human disease that we propose as novel candidates supported by embryonic lethality of their orthologs or other lines of evidence (MS4A7, SERPINA11, FCRL4, MYBPHL, PRPF19, VPS13D, KIAA1109, MOCS3, SVOPL, FEN1, HSPB11, KIF19, and EXOC3L2). Conclusion: Our results suggest that molecular autopsy in pregnancy losses is a practical and high-yield alternative to traditional autopsy, and an opportunity for bringing precision medicine to the clinical practice of perinatology.
