AN ANALYSIS OF STRUCTURAL INSTANCES OF LOW-COMPLEXITY SEQUENCE SEGMENTS

Amino acid sequence databases contain many low complexity, composition ally biased sequence segments. However, only a limited number of relat ively short instances of these segments occur in proteins of known str ucture, An analysis is presented of structural instances of these low complexity sequence segments in the Brookhaven Protein Data Bank with regard to preferences for sequence composition, secondary structural c onformation and the local atomic environment, The complexity varies al most linearly with segment length, reflecting the absence of very long , low complexity segments in the structural database. The low complexi ty segments identified are not disordered and have temperature factors which are generally the same as the rest of the protein. It is observ ed that these segments are predominantly exposed and either helical or coiled, in excess of what would be expected by chance, Secondary stru cture prediction methods perform well in correctly predicting those lo w complexity segments which are helical but poorly in correctly predic ting segments that are strands.