Chunk #46 — Methods — Statistical methods. — Deep learning annotations from DeepSEA and Basenji.

Source

Improving the trans-ancestry portability of polygenic risk scores by prioritizing variants in predicted cell-type-specific regulatory elements.

Embedded

yes

Text

center 200 bp of the 1-kb sequence falls in our provided TF ChIP–seq peaks, this sequence is labeled with a 1, otherwise it is 0. DeepSEA accurately predicted TF binding, average area under the receiver operating characteristic curve = 0.93, s.e.m. = 0.007; training was performed on chromosomes 1–5 and 10–22, testing was performed on chromosomes 8–9 and validation was performed on chromosomes 6–7.