RESUMO
Predicting protein-protein interactions (PPI) represents an important challenge in structural bioinformatics. Current computational methods display different degrees of accuracy when predicting these interactions. Different factors were proposed to help improve these predictions, including choosing the proper descriptors of proteins to represent these interactions, among others. In the current work, we provide a representative protein structure that is amenable to PPI classification using machine learning approaches, referred to as residue cluster classes. Through sampling and optimization, we identified the best algorithm-parameter pair to classify PPI from more than 360 different training sets. We tested these classifiers against PPI datasets that were not included in the training set but shared sequence similarity with proteins in the training set to reproduce the situation of most proteins sharing sequence similarity with others. We identified a model with almost no PPI error (96-99% of correctly classified instances) and showed that residue cluster classes of protein pairs displayed a distinct pattern between positive and negative protein interactions. Our results indicated that residue cluster classes are structural features relevant to model PPI and provide a novel tool to mathematically model the protein structure/function relationship.
Assuntos
Inteligência Artificial , Biologia Computacional/métodos , Bases de Dados de Proteínas/estatística & dados numéricos , Aprendizado de Máquina , Mapeamento de Interação de Proteínas/métodos , Proteínas/química , Algoritmos , Análise por Conglomerados , Análise de Sequência de Proteína/métodosRESUMO
BACKGROUND AND OBJETIVE: Human papilloma virus (HPV) is one of the main risk factors associated with the development of cervical cancer and its precursor lesions. It has been reported that HPV16 and 18 types cover approximately 70% of cervical cancer worldwide; however, significant variation in percentages of HPV infections could be related to specific populations. MATERIALS AND METHODS: Purified DNA of 67 cervical samples were analyzed by Linear Array® HPV genotyping kit. These analyzed samples correspond to 19 cervical tumors, 15 high-grade squamous intraepithelial lesions, 20 low-grade squamous intraepithelial lesions, and 13 cervical samples without injury were studied, all of them previously diagnosed. RESULTS: In general, 16 different HPV types were found with differences in their frequencies, cervical invasive cancer being the richest in HPV sequences, followed by the low-grade squamous intraepithelial lesions and then high-grade lesions. HPV16 was the most frequently distributed type in neoplastic lesions of the cervix, followed by the HPV52, suggesting viral type variability, probably associated to the geographical region studied. CONCLUSIONS: The results could indicate variability in HPV presence in Mexico, underlining the important role for HPV52 among others in the Mexican population. This would also potentially have an impact on the current anti-HPV vaccination schemes.