Desenvolvimento de um pipeline de bioinformática para análise de redes de co-expressão gênica com dados públicos de sequenciamento: estudo de caso com Spodoptera frugiperda (J. E. Smith, 1797)

Abstract

As the costs associated with nucleotide sequencing have decreased, the quantity of data available from RNA-seq work in public databases has increased exponentially. However, there are still few studies that exploit these resources, especially when considering integrative approaches such as predicting gene co-expression networks. In light of the aforementioned considerations, this work proposed the development of a Bioinformatics pipeline for obtaining, filtering and analyzing RNA-seq data publicly available in the NCBI SRA database for the prediction of gene co-expression networks, using the Spodoptera frugiperda species as a model, given its relevance as an agricultural pest with a major impact on Brazilian agriculture. Through this pipeline, 839 datasets were obtained from the SRA database and subjected to a series of filters to remove sequences and datasets of low quality or presenting risks of contamination, after which 641 datasets were selected. In addition to these filters, 339 datasets without metadata informing the tissue or life stage of the insect sampled had to be removed from the analysis. Finally, 10 conditions were selected because they contained more than 12 samples each, resulting in a total of 276 samples actually being incorporated into the co-expression analyses. In the WGCNA analyses, seven co-expression modules were predicted, including a module composed of genes most expressed on average in the gut and whole body of S. frugiperda larvae. Our results demonstrate that it is possible to use public RNA sequencing data to predict co-expression networks and propose a pipeline for filtering sequences, selecting samples and analyzing these data. Finally, we emphasize the importance of making these datasets available and fully describing them in public databases, such as SRA, for their use in integrative approaches such as the one used in this work.