Alex Chenchik, Cellecta’s founder and Scientific Director, and Paul Diehl, Director of Business Development, attended the AACR-NCI Systems Biology : Confronting the Complexity of Cancer Conference in San Diego last week and presented two posters: Identification and Analysis of Essential Genes in Leukemic Cell Lines Using RNAi Screening and HT Screening for Bioactive Peptides that Increase Radiation Tolerance Using a Pooled Lentiviral Peptide Scanning Library. Although the poster session was relatively short for a small meeting, both posters were received well. Most of the interest was on the Identification of Essential Genes in Leukemic Cells poster as the shRNA library screening technique described in this study is generally applicable for validating pathways, identifying genes modulating cellular responses, and finding putative drug targets and biomarkers, which were the main interests of most of the attendees.
Almost all the talks at the conference were fascinating and many somewhat overwhelming in terms of the experimental work and data analysis involved. Researchers are struggling with the challenge to integrate and make sense of new experimental results in light of the massive amounts of expression profiling, protein-protein interaction, SNP analysis, genomic amplification, and other available profiling data for the range of genes, transcripts, proteins, and metabolites in a range of cell lines, tissues, and tumors. There exist numerous tools and approaches to synthesize and apply established results with new data but, clearly, much more work is necessary before a comprehensive and holistic understanding of the biological mechanisms controlling the many forms of cancer and tumorigenesis emerges.
Highlights and shRNA Library Screens
One of the highlights of the conference included the opening address from Stephen Friend of Sage Bionetworks during which he discussed top-down approaches to integrate these disparate data and ways to integrate and organize data sharing between various groups to develop more comprehensive mechanistic models. Also, Louis Staudt from the NCI talked about identification of a novel drug target for B-cell lymphoma using in vivo screening of mice with tet-regulated shRNA expression libraries. In another session, Michael Hemann from the MIT Koch Institute, used in vivo shRNA screening to investigate the mechanisms of various inhibitor drugs in vivo in the B-ALL Burkett’s lymphoma mouse model and, after starting with a pooled 2,200 shRNA expression library, was able to predict the mechanism of action for various compounds with just an 8-shRNA signature.
Unrelated to shRNA profiling, Jennifer Pietenpol from the Vanderbilt-Ingram Cancer Center gave an exceptional presentation of her thorough work genetically grouping, characterizing, and identifying driver genes in triple-negative breast cancers (ER-, progesterone-, and PR-) without HER2 amplification. In the same session, Valerie Weaver from UCSF discussed the often overlooked role of interactions between tumors and extracellular matrix proteins, and how this contributes to structural changes that lead to breast cancer tumor aggressiveness and pathology. In a different session, Ernest Fraenkel from the MIT presented work showing how information from protein-protein assays and expression profiles can be integrated using a Steiner Tree algorithm to work out functionally coherent pathways and aberrant interactions in EGFR networks in glioblastoma. Also, Todd Golub from the Broad Institute distilled vast amounts of expression profiles down to a list of 1,000 “landmark genes” that can be used to infer the expression levels of almost all other genes, so a complete expression profile can be generated that’s over 80% accurate just by measuring levels of 5% of the transcripts.
Software Tools and Resources
From a very practical perspective, many of the presenters highlighted numerous web-based software and database resources to assist in building, dissecting, and analyzing pathways. Paul Spellman from the Lawrence Berkeley National Laboratory discussed on-going work of The Cancer Genome Atlas (TCGA) Project that uses HT sequencing to genotype many dozens of cancer-derived cells lines across multiple cancer types. Andrea Califano from Columbia University discussed the use of the Master Regulator Analysis module of the Genomics Workbench ge(Workbench) to eludicate mechanisms of aberrant signal transduction in various cancer sub-types. Douglas Lauffenburger, also from MIT, discussed several cell network modeling approaches from simple relational to more structured logical and mechanistic with examples of how they could be used to analyze various type of data and pathways.
Other presenters referenced a variety of quite sophisticated tools and resources for cancer research and pathway analysis, links to some of these are listed below:
Paradigm—Pathway Analysis Software (UCSC)
ICGC— International Cancer Genome Consortium
ENCODE—Encyclopedia of DNA Elements (UCSC)
COSMIC—Catalog of Somatic Mutations in Cell Cancer (Sanger Center)
Biology Workbench— Database searching, analysis, and modeling tools (UCSC)
cBio Cancer Genomics Portal—Access to large-scale genomic cancer sets (MSKCC)
dbGap—Database of Genotypes and Phenotypes (NCBI)
ARACNe— Algorithm for the Reconstruction of Accurate Cellular Networks (Columbia)
MINDy— Modulator Inference by Network Dynamics to module modulator gene interference of a network (Columbia)
NetPhorest—Analysis for phosphorylation-dependent signaling motifs
NetWorKIN—Predicting in vivo kinase-substrate relationships
mFINDER—Network motifs detection tool (Weizmann Institute)
PTMScout— Analyze mass spec and other protein data for post-translational proteomics modifications