UC San Diego

Cellular behavior is dictated by numerous factors including cell-type and environmental-dependent gene expression as well as somatically acquired mutations in cancer. The goals of my thesis research are to employ high throughput techniques and informatics to assay gene function in the regulation of disease-relevant phenotypes. First, we developed a screening platform to combine image-based readouts with the easy-to-use pooled CRISPR knockout approach. Pooled CRISPR methods allow for the genetic knock-out of tens of thousands of genes in a single pool of cells, but are limited thus far to bulk-selected phenotypes including proliferation rates and FACS-based screening. In this work, we adapted a microraft array platform and developed automated imaging and analysis programs to screen over 1,000 RNA binding proteins (RBPs) for their role in regulating the abundance of cytoplasmic stress granules, a cellular phenotype with implications in neurodegeneration. Second, to understand the role of somatic mutations in splicing factor (SF) proteins to drive cancer phenotypes, we developed a human isogenic induced pluripotent stem cell (iPSC) model of the canonical \emph{\U2AF1 S34F} and \emph{SRSF2 P95L} mutations. We found that iPSC-derived hematopoietic stem/progenitor cells (iPSC-HSPCs) have a reduced ability to terminally differentiate into mature blood cells, thus reproducing the cancer phenotype found in patients with Myelodysplastic Syndromes (MDS). An integrative RNA-Seq and eCLIP analysis these cells revealed an that alternative splicing of \emph{GNAS} is a direct target of mutant SFs. Functional studies confirmed that expression of the cancer-associated isoform, \emph{GNAS-L} reproduced our disease phenotype in normal cells, and correction of SF-mutant cells to \emph{GNAS-S} rescued the disease phenotype back to normal levels. \emph{GNAS} is a major regulator of cell signaling, and also contains point mutations in a subset of MDS patients. Mutational analysis of patient data and biochemical work on purified forms of this protein suggest usage of \emph{GNAS-L} in addition to the \emph{GNAS R201H} point mutation cooperate to induce overactivation of G-protein signaling in cancer cells. This work has identified a novel cellular function of \emph{GNAS} isoform regulation, and opened up the door for G protein signaling as a potential therapeutic avenue for SF-mutant MDS.