Table 2.

Selected features

Feature	Annotation
is_HLA	Whether the protein containing the SAP belongs to HLA family [18].
nor_diff_freq	Normalized difference between mutant-type residue frequency and wild-type residue frequency.
DAYM780301	Log odds matrix for 250 PAMs [27]. The value between two amino acids shows how often one amino acid replaces another one in evolution.
FEND850101	Structure-Genetic matrix [28]. This matrix takes into account of the structural similarities of amino acids and the genetic code.
ZHAC000105	Environment-dependent residue contact energies [29]. The residue contact energies in different structural environment.
HENS920103	BLOSUM80 substitution matrix [30]. The value between two amino acids is defined based on the log likelihood of one amino acid substitutes the other by chance in sequence alignment.
NAKH900106	Normalized composition from animal [31]. Normalized residue composition calculated from animal mitochondrial proteins.
metal_seq_neibor	The sequence distance between the SAP site and its nearest residue holding the functional site with Feature Key of METAL [18].
MIYS850103	Quasichemical energy of interactions in an average buried environment [32].
modres_seq_neibor	The sequence distance between the SAP site and its nearest residue holding the functional site with Feature Key of MOD_RES [18].