Authors:
Majid Masso
and
Iosif I. Vaisman
Affiliation:
George Mason University, United States
Keyword(s):
Delaunay tessellation, Statistical potential, Computational mutagenesis, Structure-function relationships, Random forest supervised classification, Prediction.
Related
Ontology
Subjects/Areas/Topics:
Bioinformatics
;
Biomedical Engineering
;
Data Mining and Machine Learning
;
Immuno- and Chemo-Informatics
;
Pattern Recognition, Clustering and Classification
;
Structural Bioinformatics
;
Structural Variations
Abstract:
The signaling molecule human interleukin-3 (IL-3) is responsible for promoting the growth of a wide range of hematopoietic cell lineages in the bone marrow. In this study, we apply an in silico mutagenesis technique to investigate the effects of single amino acid substitutions in the IL-3 protein on cell proliferation activity. The computational mutagenesis, which utilizes the IL-3 protein structure as well as a knowledge-based, four-body statistical potential, empirically quantifies environmental perturbations at the mutated residue position in IL-3 and at all neighboring positions in the folded structure. In particular, mutated position perturbation scores alone are capable of characterizing IL-3 residues grouped by physicochemical, functional, or structural properties. Additionally, these scores elucidate an IL-3 structure–function relationship based on a collection of 630 single residue replacements for which activity changes were experimentally measured. A random forest classifi
er trained on this dataset of experimental mutants, whose respective feature vectors include environmental changes at the mutated position and at six nearest neighbors in the IL-3 structure, achieves 80% accuracy and outperforms related state-of-the-art methods.
(More)