ANDIS is an atomic angle- and distance-dependent statistical potential devoted for protein structure quality assessment. Five angles are calculated for each atom-pair with distance < dcut (dcut is adjustable from 7.0 to 15.0 Angstrom) and residue separation ≥ 7 in protein structure. The observed frequencies of atom-pair distances and corresponding angles are converted to a combined statistical potential based on the inverse Boltzmann law. ANDIS naturally integrates the atomic orientation-dependent and distance-dependent interactions. We benchmarked ANDIS with a comprehensive list of publicly available statistical potentials. ANDIS demonstrate a significantly better ability of native recognition than the other tested potentials.  

ANDIS Resource  

 Download ANDIS standalone program: ANDIS.zip (95MB)

 Download non-redundant PDB dataset (3519 experimental structures with Identity cutoff =20%, Resolution cutoff =2.0 angstroms, R-factor cutoff =0.25): PDBdataset.zip (115MB)

 Download 175 CASP10-13 decoy sets: CASP10-13.zip (354MB)

Reference:

Zhongwang Yu, Yuangen Yao, Haiyou Deng*, Ming Yi*. ANDIS: an atomic angle- and distance-dependent statistical potential for protein structure quality assessment. BMC Bioinformatics, (2019) 20:299 [PDF]