CONICET | Buscador de Institutos y Recursos Humanos

Prognostic molecular biomarkers provide information, in a very confident way, regarding the patient outcome based on gene/protein expression profiles derived from small tissue samples. In addition, predictive molecular biomarkers provide information regarding patientresponse to therapy. The main aim of the platform is helping the user to discover and evaluate novel prognostic / predictive biomarkers based on gene expression signatures. Several concepts around the expression profile analysis (gene, platform, experiment, gene signature) were modeled in the platform. Many instances of those concepts will beinvolved in the process of generating and evaluating biomarkers. The platform provides tools for simplifying users to create, edit, import, visualize in a convenient way, persist and export the instances of those modeled concepts. The main platform input is gene signatures and experiments data. During the process of discovering new signatures, the user will be responsible of selecting the gene signature providers, the gene signature filters, and the algorithm responsible of picking up thebiomarker indicator genes. The result will be a set of genes that will be used to define the indicator of the new signatures. New algorithms can be added easily to the platform. Gene expression data, provided by experiments incorporated in the platform, are used to cluster the samples of the experiments. This clustering will let the platform validate thesignificance statistic level of the biomarker using the clinical information of the same experiment. The current version of the user interface is a prototype implemented as a desktop application but we are designing and planning the web version. Figure 1 shows the platform desktop application and a Kaplan-Meier survival curve used as one of the way fordoing the validation.

enviar mensaje