Our Technology

Context-Independent Screening
Multidimensional Variable Dose Analysis
AI-powered results

Drosophila fruit fly:  Aldona Griskeviciene / licensed from Shutterstock.com  |  Silhoettes:  Pretty Vectors / licensed from Shutterstock.com  |  Mouse:  miniwide / licensed from Shutterstock.com

VDA schematic: Sierzputowska, K., Baxter, C. R., & Housden, B. E. (2018). Variable Dose Analysis: A Novel High-throughput RNAi Screening Method for Drosophila Cells. Bio-protocol, 8(24), e3112-e3112.

Random forest model schematic:  Derivative of “Random Forest Simplified” by Venkata Jagannath, used under CC-BY-SA-4.0

Problems with Genetic Interactions

Genetic interaction (GI) screening is already a powerful tool in drug discovery and can be used to identify drug-targets, predict drug synergies, determine mechanisms of action and predict drug resistance pathways (Chan & Giaccia, 2011). Despite the potential power of this approach, it can currently only be applied on a small scale due to a lack of sufficiently high-throughput screening technologies. GI screens are also generally noisy, making identification of good drug-targets challenging. This means that unbiased mapping of genetic interactions isn’t yet possible and the potential of this approach hasn’t yet been realised. In addition, identified drug-targets often fail to translate between model systems and into human patients. A major cause of cancer drug failure is context-dependency, where a drug has different effects in each model system (Ryan et al., 2018).

The Quest Solution

Our screening platform uses a combination of innovative technologies to address these issues. First, we use cross-species genetic screening to identify candidates that have context-independent effects and therefore a greater chance of successful translation into humans. Second, we have developed a novel method called Variable Dose Analysis (VDA), that reduces screen noise and allows identification of a wider range of candidate targets. Third, a variant of the VDA method called multidimensional VDA (mVDA) will allow screens to be performed that are orders of magnitude larger than is currently possible. Finally, artificial intelligence increases screen scale and improves the quality and reliability of data arising from our screening platform. Together, these technologies result in a transformative drug-target screening platform that we call the GenPleX platform.

More Information

Current screening methods often result in low reproducibility between datasets and many years of work are required to identify the most promising and robust drug-targets for drug development. To overcome this issue, we’ve established and validated methods for drug-target screens in fruit fly cells. Fly cells have several advantages over mammalian cell culture systems for genetic screens. Their use in effect applies an evolutionary filter to identify drug-targets with consistent effects between different genetic backgrounds. Specifically, if a target is found to be effective in both fly and human cells, it is likely to be effective in all model systems, tissue cell lines and in a diverse range of human patients, thereby facilitating transfer of identified targets to clinical use. This cross-species screening approach therefore results in context-independent targets that have a high chance of successful transfer to use in human patients. This is particularly important for RAS-associated cancers where candidates often produce differing effects between models.

Inspired by our success using screens in fly cells, we established an additional technology for drug-target screens called Variable Dose Analysis (VDA). This method greatly improves the sensitivity and consistency between screens and allows identification of robust drug-targets. VDA generates a gradient of target gene knockdown efficiency, with the relative level of knockdown in each individual cell marked by expression of a fluorescent protein. Cell viability measurement can then be measured over a broad range of target gene knockdown efficiencies in a single population of cells, allowing detection of targets involving essential genes. This is vitally important for drug-target discovery because candidate drug-targets are enriched amongst essential genes. Indeed, the majority of current anti-cancer drug therapies rely on partial inhibition of essential genes. Critically, other technologies are likely to miss this type of drug-target because strong disruption of an essential gene is lethal to both diseased and healthy cells and therefore selective effects on diseased cells cannot be detected.

By contrast, the VDA method is specifically designed to detect this class of drug-target. Compared to standard methods, this approach shows that signal-to-noise ratio is improved 12-fold and detection of positive controls is improved over five-fold. VDA is therefore a high-sensitivity method to detect drug-targets involving either essential or non-essential genes.

We applied this method to identify additional candidate drug-targets for the treatment of TSC tumours and identified four existing drugs against distinct targets that may be repurposed for this disease. All of these drugs were successfully validated in multiple different systems, demonstrating the robust nature of VDA screens. Three of these candidate drugs had been missed by all previous screens for TSC drug-targets performed using other screening methods, illustrating the power of the VDA method to detect robust candidate drug-targets. VDA represents an optimal method for the identification of genetic interactions and our work clearly demonstrates the capability of the VDA method to robustly identify drug-targets relevant to human disease.

Large-scale genetic interaction screens are, in theory, a powerful approach to drug discovery. That said, it’s not possible to perform genome-wide genetic interaction screens with existing methods due to the large number of pairwise gene combinations that must be tested (approximately 98,000,000 combinations in flies). Only one genome-wide screen has been done so far, and this was in yeast (with a smaller genome) and took over a decade to complete, illustrating the massive workload involved (Tong et al., 2004; Costanzo et al., 2016). To overcome this barrier, we have developed a modified version of VDA, called multidimensional VDA (mVDA), that allows multiple genes to be disrupted in a single cell while maintaining an independent readout of the effects of each individual disruption. Previous screening methods are only able to assess a single gene or pair of genes in each experiment, therefore limiting their scalability. By contrast, mVDA allows the single and pairwise effects of whole groups of genes to be determined from single cells within a mixed population. This technology allows screens in which the single and pairwise phenotypes of many genes can be tested in a single experiment. This therefore represents the first technology with the potential to allow genome-scale genetic interaction screens.

AI algorithms are used for the optimisation of multiple aspects of our GenPleX platform. Firstly, they’ve enabled us to enhance the signal-to-noise ratio of our VDA screening method, increasing our data quality 12-fold over existing screening methods. Secondly, we’re beginning to incorporate the use of AI into how we design our reagent libraries. Finally, AI algorithms increase the scale at which we can perform our screens when used for our platform’s readout.

Supporting Science

Housden, B. E., Li, Z., Kelley, C., Wang, Y., Hu, Y., Valvezan, A. J., … & Perrimon, N. (2017). Improved detection of synthetic lethal interactions in Drosophila cells using variable dose analysis (VDA). Proceedings of the National Academy of Sciences, 114(50), E10755-E10762. [ link ]

Sierzputowska, K., Baxter, C. R., & Housden, B. E. (2018). Variable Dose Analysis: A Novel High-throughput RNAi Screening Method for Drosophila Cells. Bio-protocol, 8(24), e3112-e3112. [ link ]

Housden, B. E., Valvezan, A. J., Kelley, C., Sopko, R., Hu, Y., Roesel, C., … & Perrimon, N. (2015). Identification of potential drug targets for tuberous sclerosis complex by synthetic screens combining CRISPR-based knockouts with RNAi. Science Signaling, 8(393), rs9-rs9. [ link ]

Housden, B. E., Li, Z., Kelley, C., Wang, Y., Hu, Y., Valvezan, A. J., … & Perrimon, N. (2017). Improved detection of synthetic lethal interactions in Drosophila cells using variable dose analysis (VDA). Proceedings of the National Academy of Sciences, 114(50), E10755-E10762. [ link ]

Valvezan, A. J., Turner, M., Belaid, A., Lam, H. C., Miller, S. K., McNamara, M. C., … & Manning, B. D. (2017). mTORC1 couples nucleotide synthesis to nucleotide demand resulting in a targetable metabolic vulnerability. Cancer Cell, 32(5), 624-638. [ link ]

Nicholson, H. E., Tariq, Z., Housden, B. E., Jennings, R. B., Stransky, L. A., Perrimon, N., … & Kaelin, W. G. (2019). HIF-independent synthetic lethality between CDK4/6 inhibition and VHL loss across species. Science Signaling, 12(601). [ link ]

Valvezan, A. J., McNamara, M. C., Miller, S. K., Torrence, M. E., Asara, J. M., Henske, E. P., & Manning, B. D. (2020). IMPDH inhibitors for antitumor therapy in tuberous sclerosis complex. JCI Insight, 5(7). [ link ]

Parkhitko, A. A., Singh, A., Hsieh, S., Hu, Y., Binari, R., Lord, C. J., … & Perrimon, N. (2021). Cross-species identification of PIP5K1-, splicing-and ubiquitin-related pathways as potential targets for RB1-deficient cells. PLoS Genetics, 17(2), e1009354.

Tong, A. H. Y., Lesage, G., Bader, G. D., Ding, H., Xu, H., Xin, X., … & Boone, C. (2004). Global mapping of the yeast genetic interaction network. Science, 303(5659), 808-813. [ link ]

Chan, D. A., & Giaccia, A. J. (2011). Harnessing synthetic lethal interactions in anticancer drug discovery. Nature Reviews Drug Discovery, 10(5), 351-364. [ link ]

Fischer, B., Sandmann, T., Horn, T., Billmann, M., Chaudhary, V., Huber, W., & Boutros, M. (2015). A map of directional genetic interactions in a metazoan cell. Elife, 4, e05464. [ link ]

Costanzo, M., VanderSluis, B., Koch, E. N., Baryshnikova, A., Pons, C., Tan, G., … & Boone, C. (2016). A global genetic interaction network maps a wiring diagram of cellular function. Science, 353(6306). [ link ]

Han, K., Jeng, E. E., Hess, G. T., Morgens, D. W., Li, A., & Bassik, M. C. (2017). Synergistic drug combinations for cancer identified in a CRISPR screen for pairwise genetic interactions. Nature Biotechnology, 35(5), 463. [ link ]

Horlbeck, M. A., Xu, A., Wang, M., Bennett, N. K., Park, C. Y., Bogdanoff, D., … & Gilbert, L. A. (2018). Mapping the genetic landscape of human cells. Cell, 174(4), 953-967. [ link ]

Ryan, C. J., Bajrami, I., & Lord, C. J. (2018). Synthetic lethality and cancer–penetrance as the major barrier. Trends in cancer, 4(10), 671-683. [ link ]

Chai, N., Haney, M. S., Couthouis, J., Morgens, D. W., Benjamin, A., Wu, K., … & Gitler, A. D. (2020). Genome-wide synthetic lethal CRISPR screen identifies FIS1 as a genetic interactor of ALS-linked C9ORF72. Brain Research, 1728, 146601. [ link ]

Kelly, M. R., Kostyrko, K., Han, K., Mooney, N. A., Jeng, E. E., Spees, K., … & Jackson, P. K. (2020). Combined Proteomic and Genetic Interaction Mapping Reveals New RAS Effector Pathways and Susceptibilities. Cancer Discovery, 10(12), 1950-1967. [ link ]

Replogle, J. M., Norman, T. M., Xu, A., Hussmann, J. A., Chen, J., Cogan, J. Z., … & Adamson, B. (2020). Combinatorial single-cell CRISPR screens by direct guide RNA capture and targeted sequencing. Nature Biotechnology, 38(8), 954-961. [ link ]

DeWeirdt, P. C., Sanson, K. R., Sangree, A. K., Hegde, M., Hanna, R. E., Feeley, M. N., … & Doench, J. G. (2021). Optimization of AsCas12a for combinatorial genetic screens in human cells. Nature Biotechnology, 39(1), 94-104. [ link ]

Shimada, K., Bachman, J. A., Muhlich, J. L., & Mitchison, T. J. (2021). shinyDepMap, a tool to identify targetable cancer genes and their functional connections from Cancer Dependency Map data. Elife, 10, e57116. [ link ]

Sun, X., Wang, Z., Chen, X., & Shen, K. (2021). CRISPR-cas9 Screening Identified Lethal Genes Enriched in Cell Cycle Pathway and of Prognosis Significance in Breast Cancer. Frontiers in Cell and Developmental Biology, 9, 456. [ link ]