Table 1 Data features extracted by PEAS from ATAC-seq bam files and genomic sequence categorized by feature type.
From: A neural network based model effectively predicts enhancers from clinical ATAC-seq samples
Feature Type | Feature Label (24) |
|---|---|
ATAC-Seq Peak Driven (n = 5) | Peak score (MACS) |
Peak length (MACS) | |
Fold change | |
Summit pileup | |
Summit center distance | |
ATAC-seq Insert/Cut Driven (n = 10) | # of all inserts |
# of inserts (0,50] | |
# of inserts (50, 150] | |
# of inserts (150, 300] | |
# of inserts (300, 500] | |
# of inserts (500,) | |
Insert size (mean) | |
Long/short insert ratio | |
# of cuts within peak | |
# of overrepresented cuts | |
Sequence Driven (n = 3) | Conservation (mean) |
GC% (HOMER) | |
CpG% (HOMER) | |
Motif Driven (n = 4) | # of CTCF motifs |
% of known motifs present | |
% of denovo motifs present | |
Genomic Location Driven (n = 3) | Annotation (HOMER) |
Distance to TSS | |
Gene type (HOMER) |
