Back

De novo variants in population constrained fetal brain enhancers and intellectual disability

De Vas, M. G.; Garstang, M. G.; Joshi, S. S.; Khan, T.; Atla, G.; Parry, D.; Moore, D.; Cebola, I.; Zhang, S.; Cui, W.; Lampe, A. K.; Lam, W. W.; FitzPatrick, D. R.; Ferrer, J.; Madapura, P. M.; Atanur, S. S.

2019-09-08 genomics
10.1101/621029 bioRxiv
Show abstract

PurposeThe genetic aetiology of a major fraction of patients with intellectual disability (ID) remains unknown. De novo mutations (DNMs) in protein-coding genes explain up to 40% of cases, but the potential role of regulatory DNMs is still poorly understood.\n\nMethodsWe sequenced 70 whole genomes from 24 ID probands and their unaffected parents and analyzed 30 previously sequenced genomes from exome-negative ID probands.\n\nResultsWe found that DNVs were selectively enriched in fetal brain-specific enhancers that show purifying selection in human population. DNV containing enhancers were associated with genes that show preferential expression in the pre-frontal cortex, have been previously implicated in ID or related disorders, and exhibit intolerance to loss of function variants. DNVs from ID probands preferentially disrupted putative binding sites of neuronal transcription factors, as compared to DNVs from healthy individuals and most showed allele-specific enhancer activity. In addition, we identified recurrently mutated enhancer clusters that regulate genes involved in nervous system development (CSMD1, OLFM1 and POU3F3). Moreover, CRISPR-based perturbation of a DNV-containing enhancer caused CSMD1 overexpression and abnormal expression of neurodevelopmental regulators.\n\nConclusionOur results, therefore, provide new evidence to indicate that DNVs in constrained fetal brain-specific enhancers play a role in the etiology of ID.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.