Chemical-genetic approaches offer the potential for unbiased functional annotation of chemical libraries. Mutations can alter the response of cells in the presence of a compound, revealing chemical-genetic interactions that can elucidate a compound's mode of action. We developed a highly parallel, unbiased yeast chemical-genetic screening system involving three key components. First, in a drug-sensitive genetic background, we constructed an optimized diagnostic mutant collection that is predictive for all major yeast biological processes. Second, we implemented a multiplexed (768-plex) barcode-sequencing protocol, enabling the assembly of thousands of chemical-genetic profiles. Finally, based on comparison of the chemical-genetic profiles with a compendium of genome-wide genetic interaction profiles, we predicted compound functionality. Applying this high-throughput approach, we screened seven different compound libraries and annotated their functional diversity. We further validated biological process predictions, prioritized a diverse set of compounds, and identified compounds that appear to have dual modes of action.
Access optionsAccess options
Subscribe to Journal
Get full journal access for 1 year
only $15.58 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
This work was supported by RIKEN Strategic Programs for R&D. J.S.P. and S.C.L. were funded by a RIKEN Foreign Postdoctoral Fellowship. S.W.S. is supported by an NSF Graduate Research Fellowship (00039202), an NIH Biotechnology training grant (T32GM008347), and a one-year BICB fellowship from the University of Minnesota. H.O. is a research fellow of the Japan Society for the Promotion of Science (JSPS). R.D., J.N., E.W., and C.L.M. are supported by National Institutes of Health Grants 1R01HG005084-01A1, 1R01GM104975-01, and R01HG005853 and National Science Foundation Grant DBI 0953881. C.B. and Y.O. are supported by JSPS KAKENHI Grant Numbers 15H04483. C.B. and B.A. were supported by the Canadian Institutes of Health Research, grants FDN-143264 and FDN-143265, respectively. C.L.M., C.B., M.C., J.L., and B.A. are supported by the Canadian Institute for Advanced Research Genetic Networks Program. Y.O. is supported by Ministry of Education, Culture, Sports, Science and Technology, Japan Grant for Scientific Research 24370002 and JSPS KAKENHI Grant Numbers 15H04402. M. Yoshida is supported by JSPS KAKENHI Grant Number 26221204. A.B. is supported by a Lewis Sigler fellowship at Princeton University. G.W.B. and N.P.T. are supported by Canadian Cancer Society Research Institute impact grant 702310. K.S. is supported by The Core Research for Evolutional Science and Technology (CREST) from the Japan Science and Technology Agency (JST). We thank Astellas Pharma, Inc. (Tokyo, Japan) for their kind gift of micafungin. We thank T. Saito for help with NPDepo compound access. Sequencing was provided by RIKEN Center for Life Science Technologies, Division of Genomic Technologies, Genome Network Analysis Support Facility (GeNAS) RIKEN CLST and the University of Chicago.
Number of compounds in each collection causing >20% inhibition of growth against the yeast screening pool.
Diagnostic screening pool deletion mutant strain members.
768 multiplex tag primers.
Names, structures, doses, and chemical properties of screened compounds.
Average chemical genomic profile of 11,983 compounds.
Average chemical genomic profile of 1,522 high-confidence compounds.
Number of positive (CG score ≥+2.5), negative (CG score ≤−2.5), and all chemical genetic interactions per strain for all screened compounds.
Compounds passing our false discovery rate filter (high-confidence set) with top 3 GO process predictions, bioprocess predictions, and driver genes.
Interaction degree (positive and negative interactions) of the 275 strains from the RIKEN subset of the data also present in the S. cerevisiae genetic interaction array. Strains classified as high degree were removed to obtain “no-high-degree” process-level predictions, as were strains classified as low degree to obtain “no-low-degree” predictions.
Comparison of the compounds, and their predictions, discovered with high confidence using all, only negative, and only positive chemical-genetic interactions from the RIKEN subset of the data. Each “Negative-all” and “Positive-all” profile contained all interactions with scores below or above zero, respectively, while the pair of “Negative-equal” and “Positive-equal” profiles for each compound contained an equal number of negative and positive interactions, respectively, above a score magnitude cutoff of 1.
Growth inhibition by 24 bioactive compounds (10 μg/mL) of the screening pool versus the gtr1Δ deletion mutant (MATα pdr1Δ::natMX pdr3Δ::KI.URA3 snq2Δ::KI.LEU2 gtr1Δ::kanMX can1Δ::STE2pr-Sp_his5 lyp1Δ his3Δ1 leu2Δ0 ura3Δ0 met15Δ0).
Bioprocesses and associated GO annotation for compounds of the NIH-CC.
Functional enrichments and known targets of the GSK kinase inhibitor collection.
Cellular compartments with enrichment or depletion of targets across compound collections.
Functional enrichments among substructures. “All compounds with substructure” shows the compounds containing the substructure for that given row, and “hit compounds with substructure” shows the subset of those compounds that is annotated to any of the GO terms in that row. Substructures can be visualized directly at mosaic.cs.umn.edu/RIKEN_substructures.html or indirectly by copy-and-pasting the semicolon-delimited compound sets into the search bar at mosaic.cs.umn.edu.
Validation of predicted biological process targets of 67 compounds using cell cycle analysis.
Validation of cell wall–targeting compounds.
Prioritized functional diversity set from public collections screened.
Prioritized compound sets for 17 bioprocesses. Each high-confidence compound was placed into the group corresponding to its top bioprocess prediction in Supplementary Table 7.
Number of high confidence compounds per bioprocess, and enrichments/depletions across the tested compound collections.
Proportional differences in GO annotations versus background expectation for each library/group of compounds.
Mapping of Lee et al. major chemical-genetic signatures to bioprocesses in this study. Pearson correlation coefficients (PCCs) > 0.3 between Lee et al. compounds and high confidence compounds (this study) were mapped to major chemical-genetic signatures (Lee et al.) or bioprocesses (this study) to generate a mapping from each major chemical-genetic signature to the bioprocesses that contained similar compounds. Only the most confident mapping, as determined by relative confidence (the product of the fractions of compounds in each set that contribute to the similarities) is shown.
About this article
Nature Chemical Biology (2017)