Abstract In 2018, the FDA set a record of 59 new drug approvals. However, the percentage of first-in-class drugs to date has not exceeded 50% since 2015. The advent of innovative and orthogonal technologies focused on biological and phenotypic outcomes provides opportunity for identification of novel disease specific target identification for drug discovery programs. The Interrogative Biology® platform deconstructs the drug discovery paradigm by comprehensive molecular capture of human derived disease models. Using high-throughput multi-omic (proteomic, lipidomic, metabolomic) profiling and the causal inference-based AI tool, bAIcis®, the BERG Interrogative Biology® platform, identifies new therapeutic targets and biomarkers, as well as provides de novo inference to mechanism of action. BERG's pan-cancer models were built using multiple normal/cancer model systems covering broad etiologies and mutational profiles. A recently published study, Behan FM et al. (Nature, 2019 Apr 10) combined 941 CRISPR - CAS9 (CC9) drop-out screens, from 324 human cancer cell lines, with genetic information to generate a list of 628 priority targets across 19 different tissues. The congruent gene-phenotype linkage of targets identified from BERG's multi-omics pan-cancer profiles to the gene based CC9 was identified from an in-silico data-set overlay. Approximately 80% of the priority targets identified in the CC9 study were found to be present in BERG's pan-cancer networks identified using Interrogative Biology® platform. Overlay of CC9 targets on BERG's pan-cancer models identified 159 (protein based) priority candidate targets. Of the top 46 targets, 15 had the same protein, gene, and cancer types in both BERG and CC9 study. These protein targets demonstrated dependencies to well defined oncogenes and represent “high confidence” targets for immediate next steps in development. The remaining 13 targets, highly ranked by both Behan's and BERG's methods, are “low risk” pan-cancer targets. A set of 113 other potential targets were ranked in the CC9 study, but not by BERG's approach, represent low priority targets, requiring additional validation prior to hit finding screens and druggability assessment. The identification of a high percentage of protein targets, from phenotypic biology-based pan-cancer models in a gene based CRISPr-CAS9 screen, exemplifies the robustness of multi-omics profiling and Bayesian AI analytics in identifying de novo therapeutic targets in a reliable and unbiased manner, for rapid validation and deployment into drug discovery programs. Citation Format: Leonardo O. Rodrigues, Lixia Zhang, Poornima Tekumalla, Stephane Gesta, Vivek K. Vishnudas, Michael A. Kiebish, Niven R. Narain, Rangaprasad Sarangarajan. Benchmarking targets from cancer models using causal inference based drug-target and phenotype identification (Interrogative Biology®) cross-validates “high-priority” targets identified in CRISPR-CAS9 screen [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2929.