Reconhecimento de padrões em biossequências utilizando sistema imunológico artificial
Carregando...
Data
2021-09-08
Autores
Liberato, Luiz Paulo
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Com o avanço nos estudos genômicos foi possível entender melhor a herança genética, a síntese de proteínas e as mutações que ocorrem nos seres vivos. Com o aumento na capacidade de sequenciamento do ADN (Ácido desoxirribonucleico), e seu armazenamento, torna-se possível estudos biológicos avançados. O crescimento dos dados agrega massa de conhecimento para profissionais da área de genética, no entanto, o processamento passa a ser dispendioso quando utilizados métodos determinísticos. Para garantir tempo hábil e maior precisão no processo de reconhecimento de padrões utiliza-se de métodos heurísticos, dado que métodos determinísticos inviabilizam a execução de grandes volumes de dados. Métodos heurísticos possuem a característica de buscar a melhor solução possível dentro do espaço de busca que é explorado. Dentre as heurísticas conhecidas tem-se o Sistema Imunológico Artificial (SIA) que se enquadra na categoria de métodos bioinspirados que simulam um comportamento biológico. No presente trabalho desenvolveu-se a implementação do CLONALG (Algoritmo de Seleção Clonal) da abordagem do SIA com o MMO (Modelo de Markov Oculto) como função de afinidade, afim de obter padrões estocásticos que representem informações genéticas com relevância biológica e um tempo computacional aceitável. Como resultado foi obtido um valor 50% mais relevante em termos de tempo de execução, quando comparado ao CLONALG com a função de afinidade de Hamming. Por se tratar de uma abordagem estocástica é possível armazenar os padrões com maior afinidade para processamentos futuros, e ajustes nos parâmetros do algoritmo podem ser feitos para melhorar ainda mais a qualidade dos padrões encontrados. Finalmente, também validou-se que o CLONALG com a implementação MMO foi capaz de reconhecer os mesmos padrões quando comparado a ferramentas similares.
With the advance on genomics studies was possible to know better the genetic inheritance, protein synthesis and mutations that occurs in living beings. With the increase in the DNA sequencing capacity (Deoxyribonucleic acid), and its storage, advanced biological studies are possible. The growth of data adds mass of knowledge for professionals in the field of genetics, however, processing becomes expensive when using deterministic methods. To ensure timely and greater precision in the pattern recognition process, heuristic methods are used, since deterministic methods make it impossible to execute large volumes of data. Heuristic methods have the characteristic of seeking the best possible solution within the search space that is explored. Among the known heuristics is the Artificial Immune System (AIS), which falls under the category of bioinspired methods that simulate biological behavior. In this work, the CLONALG (Clonal Selection Algorithm) of the AIS approach was implemented with HMM (Hidden Markov Model) as an affinity function, in order to obtain stochastic patterns with biological relevance and an acceptable computational time. As a result, a 50% more relevant value was obtained in terms of execution time, when compared to CLONALG with the Hamming affinity function. As this is a stochastic approach, it is possible to store the patterns with greater affinity for future processing, adjustments in the algorithm parameters can be made to further improve the quality of the patterns found. Finally, it was also validated that CLONALG with the HMM implementation was able to recognize the same patterns when compared to similar tools.
With the advance on genomics studies was possible to know better the genetic inheritance, protein synthesis and mutations that occurs in living beings. With the increase in the DNA sequencing capacity (Deoxyribonucleic acid), and its storage, advanced biological studies are possible. The growth of data adds mass of knowledge for professionals in the field of genetics, however, processing becomes expensive when using deterministic methods. To ensure timely and greater precision in the pattern recognition process, heuristic methods are used, since deterministic methods make it impossible to execute large volumes of data. Heuristic methods have the characteristic of seeking the best possible solution within the search space that is explored. Among the known heuristics is the Artificial Immune System (AIS), which falls under the category of bioinspired methods that simulate biological behavior. In this work, the CLONALG (Clonal Selection Algorithm) of the AIS approach was implemented with HMM (Hidden Markov Model) as an affinity function, in order to obtain stochastic patterns with biological relevance and an acceptable computational time. As a result, a 50% more relevant value was obtained in terms of execution time, when compared to CLONALG with the Hamming affinity function. As this is a stochastic approach, it is possible to store the patterns with greater affinity for future processing, adjustments in the algorithm parameters can be made to further improve the quality of the patterns found. Finally, it was also validated that CLONALG with the HMM implementation was able to recognize the same patterns when compared to similar tools.
Descrição
Palavras-chave
Bioinformática, Sistema imunológico artificial, Reconhecimento de padrões, Bioinformatics, Artificial immune system, Pattern recognition