OCT 06, 2016 7:30 AM PDT

Keynote Speaker: Reference Component Analysis of Single Cell Transcriptomes Reveals Cellular Heterogeneity in Colorectal Cancer

  • Associate Director/Integrative Genomics and Group Leader, Genome Institute of Singapore
      Shyam Prabhakar obtained a B.Tech in Electronics and Communications Engineering from IIT-Madras and a PhD in Applied Physics from Stanford University. He was the sole recipient of the 2001 American Physical Society thesis award in Beam Physics. As a postdoctoral fellow under Eddy Rubin at the Lawrence Berkeley National Laboratory, he led seminal studies of human evolution through gene regulatory changes. In 2007 he joined the Genome Institute of Singapore, where his group uses single-cell RNA-seq, cohort-scale ChIP-seq and other high-throughput assays to uncover molecular mechanisms and diagnostic or prognostic markers of human diseases. In parallel, the group develops computational algorithms for deriving biological insights from functional genomics data.


    Intra-tumor heterogeneity is a major obstacle to cancer treatment. Existing single-cell studies of intra-tumor heterogeneity have largely focused on DNA mutations; functional heterogeneity is thus less well understood. We performed an unbiased analysis of functional diversity in colorectal cancer cells and their microenvironment using RNA-seq profiling of over 1,500 unsorted single cells from 11 primary tumors and matched normal mucosa (NM). To robustly interpret single-cell transcriptomes, we developed novel algorithms for normalizing expression estimates (pQ), clustering cells (RCA) and identifying differentially expressed genes (NODES). RCA identified 6 major cell types and multiple subtypes within colorectal samples. Single-cell differential expression analysis yielded results that were substantially different from bulk-sample analysis. Notably, epithelial-mesenchymal transition (EMT) genes were upregulated in tumor fibroblasts, but not in cancer cells. Though cancer cells generally lay on a continuum of transcriptomic states, a small "tail" of cells showed exceptionally high stemness signatures. Importantly, colorectal tumors previously assigned to a single subtype based on bulk transcriptomics could be divided into subgroups with divergent survival probability based on single-cell signatures, thus underscoring the prognostic value of our approach. Going forward, we see single-cell transcriptomics becoming an essential tool for cancer biology, biomarker discovery and personalized oncology.

    Show Resources
    Show Resources