Table 1.

Properties of involved computational tools

Tool	Modeling approach	Model feature	Output	Website	Refs.
FunSeq2	Knowledge-based	Evolutionary parameters; ENCODE summaries; PWMs; likely target genes; biological networks; recurrent elements across cancer samples	Cancer driver mutations	http://funseq2.gersteinlab.org/	[14]
CADD	Supervised learning	Evolutionary parameters; ENCODE summaries; population frequencies; transcript information; protein-level scores	Functional variants	https://cadd.gs.washington.edu/	[16]
GWAVA	Supervised learning	Evolutionary parameters; ENCODE summaries; population frequencies	Disease-related variants	https://www.sanger.ac.uk/science/tools/gwava	[17]
Eigen	Unsupervised learning	Evolutionary parameters; ENCODE summaries; population frequencies	Functional variants	http://www.columbia.edu/∼ii2135/eigen.html	[18]
DeepSEA	Supervised learning (DL)	Local sequences; evolutionary parameters	Functional variants	http://deepsea.princeton.edu/	[20]
EnsembleExpr	Ensemble-based	Including features used by DeepSEA, DeepBind, KSM, and ChromHMM	Expression-modulating variants	http://ensembleexpr.csail.mit.edu/	[21], [23], [24]
ExPecto	Supervised learning (DL)	Local sequences	Expression-modulating variants	https://hb.flatironinstitute.org/expecto/	[22]

Note: ENCODE, Encyclopedia of DNA Elements; PWM, position weight matrix; DL, deep learning; KSM, k-mer set memory.