Improved estimation of cancer dependencies from large-scale RNAi screens using model-based normalization and data integration

McFarland, James M.; Ho, Zandra V.; Kugener, Guillaume; Dempster, Joshua; Montgomery, Phillip G.; Bryan, Jordan; Krill-Burger, John M.; Green, Thomas M.; Vázquez, Francisca; Boehm, Jesse S.; Golub, Todd R.; Hahn, William C.; Root, David E.; Tsherniak, Aviad

doi:10.1101/305656

Cited by 50 publications

(80 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tumor-derived cell line models have been a cornerstone of cancer research for decades. The genomic and molecular features of over a thousand cancer cell line models have now been deeply characterized 1 , and recent efforts are systematically mapping their genetic [2][3][4] and chemical 5 vulnerabilities. These datasets are thus providing new opportunities to identify potential therapeutic targets and connect these vulnerabilities with measurable biomarkers that can be used to develop precision cancer approaches 2,5 .…”

Section: Introductionmentioning

confidence: 99%

Global computational alignment of tumor and cell line transcriptional profiles

Warren

Jones

Shibue

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Cell lines are key tools for preclinical cancer research, but it remains unclear how well they represent patient tumor samples. Identifying cell line models that best represent the features of particular tumor samples, as well as tumor types that lack in vitro model representation, remain important challenges. Gene expression has been shown to provide rich information that can be used to identify tumor subtypes, as well as predict the genetic dependencies and chemical vulnerabilities of cell lines. However, direct comparisons of tumor and cell line transcriptional profiles are complicated by systematic differences, such as the presence of immune and stromal cells in tumor samples and differences in the cancer-type composition of cell line and tumor expression datasets. To address these challenges, we developed an unsupervised alignment method (Celligner) and applied it to integrate several large-scale cell line and tumor RNA-Seq datasets. While our method aligns the majority of cell lines with tumor samples of the same cancer type, it also reveals large differences in tumor/cell line similarity across disease types. Furthermore, Celligner identifies a distinct group of several hundred cell lines from diverse lineages that present a more mesenchymal and undifferentiated transcriptional state and which exhibit distinct chemical and genetic dependencies. This method could thus be used to guide the selection of cell lines that more closely resemble patient tumors and improve the clinical translation of insights gained from cell line models.

show abstract

Section: Introductionmentioning

confidence: 99%

Global computational alignment of tumor and cell line transcriptional profiles

Warren

Jones

Shibue

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…To test whether USP15 gene expression correlates with its genetic dependency, we ranked the dependency scores calculated by Demeter2 for USP15 RNAi in CCLE lines (McFarland et al, 2018). According to DepMap (https://depmap.org), USP15 expression and dependency varied across cell lines but were not linearly correlated, nor leukemia cell lines were specifically sensitive compared to other cancers ( Fig.…”

Section: Usp15 Is Highly Expressed In Human Leukemiamentioning

confidence: 99%

USP15 deubiquitinase safeguards hematopoiesis and genome integrity in hematopoietic stem cells and leukemia cells

Berk

Lancini

Serresi

et al. 2020

Preprint

View full text Add to dashboard Cite

Altering ubiquitination by disruption of individual deubiquitinating enzymes (DUBs) has proven to affect hematopoietic stem cell (HSC) maintenance. However, comprehensive knowledge of DUB function during hematopoiesis in vivo is lacking. To accomplish this goal, we systematically inactivated DUBs in mouse hematopoietic progenitors using in vivo small hairpin RNAs (shRNAs) screens. We found that multiple DUBs may be individually required for hematopoiesis and that the ubiquitin-specific protease 15 (USP15) is particularly important for the maintenance of murine hematopoietic stem and progenitor cells in vitro and in vivo.Consistently, Usp15 knockout mice exhibited a reduced HSC pool. The defect was intrinsic to HSCs, as demonstrated by competitive repopulation assays. Importantly, USP15 is highly expressed in normal human hematopoietic cells and leukemias, and USP15 depletion in murine early progenitors and myeloid leukemia cells impaired in vitro expansion and increased genotoxic stress. Our study underscores the importance of DUBs in preserving normal hematopoiesis and uncovers USP15 as a critical DUB in safeguarding genome integrity in HSC and in leukemia cells.

show abstract

“…The DepMap website (https://depmap.org/) provides two separate pre-processed genome-wide genetic perturbation data in hundreds of cell lines, in which shRNA and CRISPR efficacy data were normalized using DEMETER2 and CERES algorithms, respectively (Meyers et al, 2017;McFarland et al, 2018). In both, genes with more negative values are considered essential in the corresponding cells.…”

Section: Computation Of a Unified 'Perturbation Score'mentioning

confidence: 99%

A tool for browsing the Cancer Dependency Map reveals functional connections between genes and helps predict the efficacy and selectivity of candidate cancer drugs

Shimada

Mitchison

2019

Preprint

View full text Add to dashboard Cite

Individual cancers rely on distinct essential genes for their survival. The Cancer Dependency Map (DepMap) is an ongoing project to uncover gene dependency in hundreds of cancer cell lines. DepMap is a powerful drug discovery tool, but can be challenging to use without professional bioinformatics assistance. We combined CRISPR and shRNA screening data from DepMap and built a non-programmer-friendly browser (https://labsyspharm.shinyapps.io/depmap) that reports, for each gene, the growth reduction that can be expected on loss of a gene or inhibition of its action (efficacy) and the selectivity of this effect across cell lines. Cluster analysis revealed proteins that work together in pathways or complexes. This tool can be used to 1) predict the efficacy and selectivity of candidate drugs; 2) identify targets for highly selective drugs; 3) identify maximally sensitive cell lines for testing a drug; 4) target hop, i.e., navigate from an undruggable protein with the desired selectively profile, such as an activated oncogene, to more druggable targets with a similar profile; and 5) identify novel pathways needed for cancer cell growth and survival. Keywords (up to 10):Cancer Dependency Map, precision medicine, essential genes, CRISPR, RNAi, synthetic lethality, ECHODOTS, cluster analysis, web tool, shiny to run the tool locally, but it will not be updated. In the following, the analysis workflow is explained ( Fig. 1B). Efficacy-Selectivity (All genes)This service allows a user to explore efficacy and selectivity. The output is split into two panels. On the left, a scatterplot displaying efficacy and selectivity of all the genes is always shown (3, the number corresponds in Fig. 1B-C). By hovering over the points on the plot with the cursor, one can identify specific genes. When a partial or full gene name is provided in a text box in the input sidebar (1), genes that are partially matched to the query as well as their efficacy and selectivity are listed on the right panel, when the "Gene list" tab is selected (4). One of the listed genes is highlighted in pink.The points on the scatter plot corresponding to the listed genes are shown in orange with the highlighted gene on the list in red. When switching to the "Lineage" tab, perturbation scores of the highlighted gene in individual cell lines are shown, stratified by lineages (5). To view the scores of other genes on the Gene list, the user selects a different gene from the drop-down menu in the Input sidebar (2).Functional clusters (essential genes only) 2,492 genes were found essential in our analysis. We clustered perturbation profiles of the genes, as described in detail below. The output consists of three panels. On top left is the same essentiality vs selectivity plot of the essential genes (9). The bottom left plot shows the similarity of perturbation score profiles across genes (10). Adjacent points in the plot are expected to be functionally connected. When one essential gene is chosen from the top left drop-down menu (6), the cluster containing the query ge...

show abstract

Improved estimation of cancer dependencies from large-scale RNAi screens using model-based normalization and data integration

Cited by 50 publications

References 27 publications

Global computational alignment of tumor and cell line transcriptional profiles

Global computational alignment of tumor and cell line transcriptional profiles

USP15 deubiquitinase safeguards hematopoiesis and genome integrity in hematopoietic stem cells and leukemia cells

A tool for browsing the Cancer Dependency Map reveals functional connections between genes and helps predict the efficacy and selectivity of candidate cancer drugs

Contact Info

Product

Resources

About