Structure-derived substitution matrices for alignment of distantly relatedsequences

Sequence alignment is a standard method to infer evolutionary, structural, and functional relationships among sequences. The quality of alignments dep ends on the substitution matrix used. Here we derive matrices based on supe rimpositions from protein pairs of similar structure, but of low or no sequ ence similarity. In a performance test the matrices are compared with 12 ot her previously published matrices. It is found that the structure-derived m atrices are applicable for comparisons of distantly related sequences. We i nvestigate the influence of evolutionary relationships of protein pairs on the alignment accuracy.