E-Locus - Institutional Repository of the University of Crete - Evolutionary models of amino acids substitutions based on their neighborhood tertiary structure

Home Evolutionary models of amino acids substitutions based on their neighborhood tertiary structure

Results - Details

[Add to Basket]

Identifier

000412711

Title

Evolutionary models of amino acids substitutions based on their neighborhood tertiary structure

Alternative Title

Εξελικτικά μοντέλα αμινοξικών υποκαταστέσεων βασιζόμενα στην γειτονική τριτοταγή δομή τους

Author

Πριμέτης, Ηλίας Ν.

Thesis advisor

Παυλίδης, Παύλος

Reviewer

Ηλιόπουλος, Ιωάννης
Τζαμαρίας, Δημήτρης

Abstract

Intra-protein interactions depend on the involved amino acids. It is assumed that the energetically favourable interactions have been preserved, while the unfavourable have been eliminated during evolution. We have used the Protein Interaction Statistics (PrInS) algorithm to statistically describe interactions between amino acids using protein structures. PrInS produces a scoring matrix to describe the frequency of amino acid interactions in the protein structures. In this project, we used structures of alpha helical membrane proteins from the RCSB PDB database. The resulting scoring matrix was converted to an amino acids distance (Euclidean, Manhattan or Pearson) matrix M, where M_ij value denotes the distance between the neighbourhoods of amino acid i and j. To test the validity of our methodology, we counted the observed number of amino acid changes in 224 alignments of homologous proteins and we correlated them with the M matrix. Assuming human protein sequences as a reference, we calculated the distance between human and 19 other species (16 primates and 3 other mammals) for all 224 proteins of our dataset, using our approach, BLOSUM62 and PAM120 amino acid substitution matrices. Outcomes were comparable, suggesting that our approach captures information about protein evolution process in a similar fashion as BLOSUM62 and PAM120. Finally, distance matrices were converted to rate matrices to calculate the likelihood of multiple alignments and the likelihood of each site in alignments.

Language

English

Subject

Computational biology

Evolution

Proteins

Structural biology

Δομική βιολογία

Εξέλιξη

Πρωτεϊνες

Υπολογιστική βιολογία

Issue date