Read on Twitter
Super excited to share our new preprint!
Using exome sequencing data from ~10k patients with severe developmental disorders (DD) from the DDD project we decipher 3 distinct mechanisms through which ‘non-coding’ variants cause MEF2C loss-of-function https://www.medrxiv.org/content/10.1101/2020.11.15.20229807v1 1/10
Using exome sequencing data from ~10k patients with severe developmental disorders (DD) from the DDD project we decipher 3 distinct mechanisms through which ‘non-coding’ variants cause MEF2C loss-of-function https://www.medrxiv.org/content/10.1101/2020.11.15.20229807v1 1/10
Two 5’UTR SNVs create upstream start codons (uAUGs) that are out-of-frame with the coding sequence (CDS) and reduce normal translation of MEF2C. This effect is dependent on the Kozak consensus of the uAUG, with even partial reduction of MEF2C leading to DD @nickquaife1 2/10
Two additional 5’UTR SNVs create uAUGs in-frame with the CDS. These variants add only a small amount of sequence to the N-terminus (3 and 9 AAs), but disrupt binding of MEF2C (a transcription factor) to DNA. These variants are found recurrently de novo in unrelated cases 3/10
Fun fact: de novo missense variants in MEF2C in DD cases also cluster at the N-terminus. Protein modelling (by @matteoferla & @carolinefwright) shows that these variants likely also cause LoF of MEF2C through disrupting DNA-binding. Take a look: https://michelanglo.sgc.ox.ac.uk/r/mef2c 4/10
Finally, two large deletions (41kb and 248kb) remove the promoter and the most 5’ exon of the 5’UTR and are predicted to abolish normal transcription of MEF2C.
All of the patients with these variants have phenotypes consistent with MEF2C haploinsufficiency 5/10
All of the patients with these variants have phenotypes consistent with MEF2C haploinsufficiency 5/10
Together, these non-coding variants comprise 23% (yes nearly a quarter!) of all likely diagnostic variants in MEF2C in DDD - not only do non-coding variants reveal cool regulatory mechanisms, they are also an important cause of severe disease! 6/10
And we can readily search for these variants in existing targeted sequencing datasets – we find that 30.7% of all 5’UTR bases are sufficiently covered (>10x) in the DDD exome sequencing data (although this is highly gene dependent) 7/10
But a caution here that the pattern does not translate across all DD haploinsufficient genes – MEF2C likely shows an enrichment of non-coding mechanisms due to multiple factors including good coverage, multiple possible high-impact variants, and high dosage sensitivity 8/10
Finally, I want to highlight that this was a huge team effort! Thank you especially to @carolinefwright, @pjrbarton, @ELaraPezzi, @nickquaife1 and Laura Ramos-Hernández (not on Twitter). You are all absolutely awesome and working with you on this was amazing fun! 9/10
We would love to hear your feedback on this work, so please let us know what you think! 10/10
