SVAI uses “collaborative research events” (CREs) to apply the power of the computational biology community to the needs of targeted patient communities.  Like a movie production company, CREs pool the talents of a diverse set of participants for a short period of time to create a well defined deliverable.  CRE deliverables increase knowledge about the disease of interest and can be leveraged in future CREs.  For example the July 2017 CRE lasted 3 days and centered on NF2.  During that time, several groups made especially significant progress.

  • The “AutoNF2 Team” extracted interaction networks from complex genomic data using “deep learning” by feeding an Autoencoder a vector of gene expressions (FPKM) to doing dimensionality reduction.  Large network weights indicated which feature was important.  Using this approach, the team postulated “miR-200a as a potential circulating marker for NF2”
  • The “C3 Team” created 7 mutational signature clusters of brain cancer using 290 glioblastomas and 286 gliomas from TCGA.  Cluster analysis suggested that drugs targeting homologous repair, e.g. PARP inhibitors, could be used to treat the NF2 patient.

The p1RCC CRE will apply these NF2 CRE methods to the p1RCC dataset.