Anorectal malformations (ARMs, congenital obstruction of the anal opening) are among the most common birth defects requiring surgical treatment (2-5/10 000 live-births) and carry significant chronic morbidity. ARMs present either as isolated or as part of the phenotypic spectrum of some chromosomal abnormalities or monogenic syndromes. The etiology is unknown. To assess the genetic contribution to ARMs, we investigated single-nucleotide polymorphisms and copy number variations (CNVs) at genome-wide scale. A total of 363 Han Chinese sporadic ARM patients and 4006 Han Chinese controls were included. Overall, we detected a 1.3-fold significant excess of rare CNVs in patients. Stratification of patients by presence/absence of other congenital anomalies showed that while syndromic ARM patients carried significantly longer rare duplications than controls (P = 0.049), non-syndromic patients were enriched with both rare deletions and duplications when compared with controls (P = 0.00031). Twelve chromosomal aberrations and 114 rare CNVs were observed in patients but not in 868 controls nor 11 943 healthy individuals from the Database of Genomic Variants. Importantly, these aberrations were observed in isolated ARM patients. Gene-based analysis revealed 79 genes interfered by CNVs in patients only. In particular, we identified a de novo DKK4 duplication. DKK4 is a member of the WNT signaling pathway which is involved in the development of the anorectal region. In mice, Wnt disruption results in ARMs. Our data suggest a role for rare CNVs not only in syndromic but also in isolated ARM patients and provide a list of plausible candidate genes for the disorder.

Anorectal malformations (ARMs) are birth defects that require surgery and carry significant chronic morbidity. Our earlier genome-wide copy number variation (CNV) study had provided a wealth of candidate loci. To find out whether these candidate loci are related to important developmental pathways, we have performed an extensive literature search coupled with the currently available bioinformatics tools. This has allowed us to assign both genic and non-genic CNVs to interrelated pathways known to govern the development of the anorectal region. We have linked 11 candidate genes to the WNT signalling pathway and 17 genes to the cytoskeletal network. Interestingly, candidate genes with similar functions are disrupted by the same type of CNV. The gene network we discovered provides evidence that rare mutations in different interrelated genes may lead to similar phenotypes, accounting for genetic heterogeneity in ARMs. Classification of patients according to the affected pathway and lesion type should eventually improve the diagnosis and the identification of common genes/molecules as therapeutic targets.