Publicação
Enhancing genomic discovery through taxonomic identification tools
| Resumo: | Metagenomics is a branch of genomics that aims to investigate complex microbial communities directly from environmental samples, without the need for laboratory cultivation. This methodol has had a significant impact across various fields, including human health, evolutionary biology, microbial ecology, agriculture, and biotechnology. However, the field faces significant challenges, including the complexity of analyzing large volumes of data, high computational resource requirements, limitations in current metagenomic tools and species identification, all of which hinder the full exploration of metagenomic data. To address these challenges, this dissertation aims to develop a metagenomic tool for taxonomic classification. To this end, a systematic review of state-of-the-art metagenomic tools was conducted, evaluating their classification algorithms and limitations. A review of taxonomic databases was also carried out, leading to the development of a standardized dataset representative of all domains of life, enabling fair comparison and validation of new methods. Based on these foundations, HYMET (Hybrid Metagenomic Tool) was introduced, an innovative classifier that combines optimized k-mer screening with a smart alignment system that dynamically switches between high-precision and efficient tools depending on the genomic characteristics of the sample. HYMET is based on the dynamic construction of a reference database specific to each sample and a robust weighted classification algorithm. The results demonstrate HYMET’s superior performance, being mutation-resistant and capable of classifying all domains with F1 scores above 0.90, while maintaining computational efficiency and processing datasets in under an hour. HYMET operates with minimal resource consumption: the main installation requires only 2.82 GB, dynamically adjusting storage usage between 5–30 GB, with efficient CPU (13.83%) and memory (10.91%) utilization. This dissertation not only advances the state-of-the-art in metagenomic classification but also lays the groundwork for future innovations, driving progress in the study of microbial life. |
|---|---|
| Autores principais: | Martins, Inês Branco |
| Assunto: | Metagenomics Taxonomic classification Microbial communities Bioinformatics Genomics Data analysis |
| Ano: | 2025 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Aveiro |
| Idioma: | inglês |
| Origem: | RIA - Repositório Institucional da Universidade de Aveiro |
| Resumo: | Metagenomics is a branch of genomics that aims to investigate complex microbial communities directly from environmental samples, without the need for laboratory cultivation. This methodol has had a significant impact across various fields, including human health, evolutionary biology, microbial ecology, agriculture, and biotechnology. However, the field faces significant challenges, including the complexity of analyzing large volumes of data, high computational resource requirements, limitations in current metagenomic tools and species identification, all of which hinder the full exploration of metagenomic data. To address these challenges, this dissertation aims to develop a metagenomic tool for taxonomic classification. To this end, a systematic review of state-of-the-art metagenomic tools was conducted, evaluating their classification algorithms and limitations. A review of taxonomic databases was also carried out, leading to the development of a standardized dataset representative of all domains of life, enabling fair comparison and validation of new methods. Based on these foundations, HYMET (Hybrid Metagenomic Tool) was introduced, an innovative classifier that combines optimized k-mer screening with a smart alignment system that dynamically switches between high-precision and efficient tools depending on the genomic characteristics of the sample. HYMET is based on the dynamic construction of a reference database specific to each sample and a robust weighted classification algorithm. The results demonstrate HYMET’s superior performance, being mutation-resistant and capable of classifying all domains with F1 scores above 0.90, while maintaining computational efficiency and processing datasets in under an hour. HYMET operates with minimal resource consumption: the main installation requires only 2.82 GB, dynamically adjusting storage usage between 5–30 GB, with efficient CPU (13.83%) and memory (10.91%) utilization. This dissertation not only advances the state-of-the-art in metagenomic classification but also lays the groundwork for future innovations, driving progress in the study of microbial life. |
|---|
Atividades financiadas
Não foram encontrados projetos financiados.