Login / Signup

BAFopathies' DNA methylation epi-signatures demonstrate diagnostic utility and functional continuum of Coffin-Siris and Nicolaides-Baraitser syndromes.

Erfan Aref-EshghiEric G BendRebecca L HoodLaila C SchenkelDeanna Alexis CarereRana ChakrabartiSandesh C S NagamaniSau Wai CheungPhilippe M CampeauChitra PrasadVictoria Mok SiuLauren BradyMark A TarnopolskyDavid J CallenA Micheil InnesSusan M WhiteWendy S MeschinoAndrew Y ShuenGuillaume ParéDennis E BulmanPeter J AinsworthHanxin LinDavid I RodenhiserRaoul C HennekamKym M BoycottCharles E SchwartzBekim Sadikovic
Published in: Nature communications (2018)
Coffin-Siris and Nicolaides-Baraitser syndromes (CSS and NCBRS) are Mendelian disorders caused by mutations in subunits of the BAF chromatin remodeling complex. We report overlapping peripheral blood DNA methylation epi-signatures in individuals with various subtypes of CSS (ARID1B, SMARCB1, and SMARCA4) and NCBRS (SMARCA2). We demonstrate that the degree of similarity in the epi-signatures of some CSS subtypes and NCBRS can be greater than that within CSS, indicating a link in the functional basis of the two syndromes. We show that chromosome 6q25 microdeletion syndrome, harboring ARID1B deletions, exhibits a similar CSS/NCBRS methylation profile. Specificity of this epi-signature was confirmed across a wide range of neurodevelopmental conditions including other chromatin remodeling and epigenetic machinery disorders. We demonstrate that a machine-learning model trained on this DNA methylation profile can resolve ambiguous clinical cases, reclassify those with variants of unknown significance, and identify previously undiagnosed subjects through targeted population screening.
Keyphrases
  • dna methylation
  • genome wide
  • copy number
  • gene expression
  • peripheral blood
  • machine learning
  • dna damage
  • artificial intelligence
  • resistance training
  • cancer therapy
  • deep learning
  • big data