Identification of biomarkers with predictive power in cancer: a perspective from the science of biomedical data and bioinformatics

In the study of cancer, gene expression profiles have great relevance since they show the activity of genes of interest in the tissue under analysis. The biotechnological advances and the sequencing cost reduction have allowed to produce large volumes of molecular data including gene expression profiles, which can be analyzed together with survival data (recurrence of a tumor or death) to obtain valuable information on the prognosis of the patient. The objective is to identify expression profiles that show association with clinically actionable characteristics, in response to a treatment or recurrence capacity of the tumor.The analysis of these large volumes of biomedical data requires computational, bioinformatic and biostatistical knowledge. The Bioplat platform allows to democratize these analyses and is especially useful for teams that have biological experience but not computational / biostatistical. It also integrates multiple sources of datasets, allows to incorporate the user’s own data and provides a curated database. It offers extension points so that computer scientists can easily incorporate new machine learning algorithms, tools or techniques.