Banca de QUALIFICAÇÃO: DANIELA COELHO BATISTA GUEDES PEREIRA

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : DANIELA COELHO BATISTA GUEDES PEREIRA
DATE: 28/02/2020
TIME: 10:00
LOCAL: BioME - PPg-Bioinfo
TITLE:

Classification of Mutations Associated with Cancer, through Machine Learning Algorithms.

KEY WORDS:

Missense Mutations, Predictors, Residue interaction Networks Machine Learning

PAGES: 54
BIG AREA: Ciências Biológicas
AREA: Biologia Geral
SUMMARY:

The large volume of single nucleotide polymorphism data now available motivates the development of methods to distinguish neutral changes from those associated with diseases such as cancer. Obtaining experimental knowledge about the possible association of variants with the disease is a laborious, time-consuming and expensive activity. A large number of in silico tools have been employed for the task of predicting pathogenicity, including PolyPhen-2, SIFT, FATHMM, MutationTaster-2, MutationAssessor, LRT, as well as optimized methods of combining tool scores, such as MetaLR and MetaSVM. Proteins have been studied as networks formed by amino acid residues and their interactions. This approach has been used for a better understanding of structure, protein function and analysis of the effects of mutations. In recent decades, consensus tools that integrate topological data from residue interaction networks together with predictor outputs have proven superior in performance over traditional tools. A computational approach to the prediction of missense mutations associated with cancer is proposed in this work. It combines protein sequence data, predictor outputs, residue interaction network data, clinical data associated with mutations and machine learning models. This integrated consensus classifier will predict whether the mutation is deleterious or not. The preliminary results (around 67% accuracy in the test dataset) reinforce our hypothesis that we can have a better classification of the missensse mutations associated with cancer if information obtained from residue interaction networks (Degree, Clustering Coefficient, Betweenness, Closeness, among others) and clinical data of the mutations are added to the predictor.

BANKING MEMBERS:
Presidente - 1513597 - JOAO PAULO MATOS SANTOS LIMA
Interno - 3083298 - RENAN CIPRIANO MOIOLI
Interna - 012.117.554-52 - THAIS GAUDENCIO DO REGO - UFPB