No tubes, reagents, microscopes. A group of researchers is producing state-of-the-art biological research using only the computer and the brain itself. It was through the technology that the team identified the genes responsible for the viral infection present in the Aedes aegypti mosquito using only data obtained from public banks.
The team reduced the number of genes specifically correlated with Dengue, Yellow Fever and Nile Fever, all transmitted by the vector, from the hundreds reported in the scientific literature to only four, eliminating the effect of the food process in the analysis. Thus, the finding will allow new forms of vector control and the production of rapid tests to detect infected mosquitoes.
The finding is detailed in the study “Meta-analysis of Aedes aegypti expression datasets: comparing virus infection and blood fed transcriptomes to identify markers of virus presence,” which will be available in open access until mid-January in the journal Frontiers in Bioengineering and Biotechnology.
In addition to researchers from the Center for Integration of Data and Knowledge for Health (Cidacs) and the Laboratory of Immunoparasitology, both from Fiocruz Bahia, researchers from USP (University of São Paulo), Federal University of Uberlândia and Federal University of Technology Paraná.
The study conducted by the team aimed to identify specific genes of the viral infection in Aedes aegypti, regardless of vector feeding. “The mosquito becomes infected during feeding. Thus, genes associated with the food process could be hiding those correlated with the infectious process, “explains USP researcher Kiyoshi Fukutani, one of the co-authors of the study. “No one has previously been able to remove this process from the analysis”, complements the researcher of Cidacs Artur Queiroz.
To reach the result, the researchers used algorithms to analyze 13,000 genes available in two public datasets from other studies that collected the desired sample: infected mosquitoes; non-infected; fed and non-fed. Comparing these groups, the researchers reached 110 genes with high correlation.
It was a Machine Learning methodology, called the Decision Tree, that helped the group further refine the results, reaching all four genes. Of these, there is a gene that alone is able to distinguish all the infected samples and, according to the study, “must play a fundamental role in discriminating viral infection.” To validate the finding, the researchers still compared the analysis with four other datasets.
Researchers Artur Queiroz and Kiyoshi Fukutani are part of the High Throughput Analysis Platform (PAH), associated with the Bioinformatics Platform of High Performance of Biological Data, of Cidacs, also headed by the study co-author and researcher at the Pablo Ramos center. “The aim of our group is to set up gene expression analysis protocols, enrichment pathways and gene interaction networks. We took data available in public databases and reanalyzed with other hypotheses, “explains Queiroz.
The big bet of the group – and also the Cidacs – is that producing new research with public data in a computational environment focused on large data analysis helps to respond more accurately and in less time great scientific questions of the field of health. “Data are available, but few reuse it,” says Fukutani, pointing out the difficulty of working with Big Data as a possible technical limitation, and the lack of interest in the scientific community to reanalyze data from other researchers. “The potential for this type of research is enormous,” he concludes.
The next step of the group is to use the same framework (set of 110 genes with high correlation) to apply in new data from the surveys with Zika. Although the potential of what can be done with what has already been published by the team is encouraging: by identifying this small group of genes, it is even possible to develop solutions that neutralize viral infection of the mosquito.
In order for the finding to leave the computational servers to the streets, with application in the control of the infections caused by Aedes aegypti, there is yet another step: that another group of researchers, this time those who wear lab coats and tubes, conduct in vitro tests. “We did all the evaluation and validation in silico [through computer simulation]. Our study is robust,” says Queiroz.