AMINO-ACID DOUBLETS AND TRIPLETS IN PROTEIN SEQUENCES - A DATABASE ANALYSIS

Sequence analysis on the aspect of amino acid doublet and triplet repe ats in 36,000 proteins comprising a total of 1,24,96,420 amino acid en tries was carried out to study the structural aspects of proteins. Any deviation (normalized) of the doublet and triplet occurrence from the ir anticipated occurrence based on the frequency of the individual ami no acid entries was attributed either to preferential selection (+ve d eviation) or to non preferential selection (-ve deviation). All the ho modoublets and homotriplets, except Ile-Ile-Ile, show positive deviati on, implying the preferential selection for homomers. Most of the hete rodoublets show marginal deviation indicating the expected distributio n. A few doublets do show high positive and high negative deviations, containing the less propensity amino acids in the repeat. Most of the triplets which contain two identical amino acids in the repeat show po sitive deviations. In this category too, the high positively deviated triplets mainly contain less propensity amino acids, substantiating th e economical usage of less propensity amino acids. A few triplets whic h contain two proline residues have large difference in observed and e xpected frequencies, in addition to high positive deviation indicating important structural role of proline residues. Most of the heterotrip lets which have three different amino acid entries in the repeat show marginal deviations. The triplets with large deviation in this categor y also are dominated by less propensity amino acids. High frequency do ublet pairs containing high propensity amino acid in the repeat (Glu-L ys, Lys-Glu), contribute structurally to alpha-helical conformation.