Ligand binding residue prediction method based on multi-sequence joint matching information

Abstract

A ligand binding residue prediction method based on multi-sequence binding information comprises the following steps: firstly, according to an input protein sequence to be subjected to ligand bindingresidue prediction and the number of binding residues, acquiring the multi-sequence binding information of protein by using an HHbits program; secondly, calculating the frequency of the same residuesappearing at the corresponding positions of the protein sequence to be predicted and the multi-sequence joint information; thirdly, according to the frequency obtained through calculation and the number of input binding residues, acquiring a pseudo-cocorrelation coefficient matrix through calculation; finally, obtaining a maximum value in the pseudo-cocorrelation coefficient matrix, and outputtingresidue information of binding the same ligand in the to-be-predicted protein sequence according to a position corresponding to the maximum value. The invention provides a ligand binding residue prediction method based on multi-sequence joint matching information, which is low in calculation cost and high in prediction precision.

Description

technical field [0001] The invention relates to the fields of bioinformatics and computer applications, in particular to a method for predicting ligand binding residues based on multiple sequence alignment information. Background technique [0002] The interaction between proteins and ligand molecules is realized through the interaction between some amino acid residues and ligand molecules. This interaction is ubiquitous and indispensable in life activities. These amino acid residues are called binding molecules. Determine residues. Therefore, the precise identification of binding residues between proteins and ligands has important guiding significance for understanding protein functions, analyzing the relationship between biomolecules, and designing new drugs. [0003] Research literature found that many methods for predicting binding residues have been proposed, such as: COACH (Yang J, RoyA, Zhang Y. Protein–ligand binding site recognition using complementary binding-spec...

AI Technical Summary

Problems solved by technology

Although existing methods can be used to predict binding residues, a large number of training data sets and machine learning algorithms are commonly used, so the calculation cost is high, and because the noise information in the training set has not received enough attention, the prediction accuracy cannot be guaranteed is optimal

[0004] In summary, the existing methods for predicting ligand-binding residues are still far from the requirements of practical applications in terms of computational cost and prediction accuracy, and urgently need to be improved.

Images