Abstract 1935P
Background
The traditional drug development process is driven by individual hypotheses, which is becoming less and less efficient. With the explosion of big cancer genomic data, the advancement of artificial intelligence, and the availability of cloud computing technologies, the era of data-driven drug development has arrived. However, there is still a large gap to put this vision into practice.
Methods
We have collected and curated high-quality multi-omics data of patient and cell line samples from (i) several large cancer genomics consortium projects, (ii) published studies, and (iii) our in-house datasets. With cutting-edge bioinformatics pipelines, computational algorithms, and AI technologies, we have developed and implemented a big data analytics platform that aims to address the challenging questions in drug development.
Results
We have developed the analytics platform, AIBERT (AI-driven Big data analytics for Effective Response of cancer Therapy), to provide innovative, translational, data-driven service for pharmaceutical companies. AIBERT consists of the well-annotated, curated multi-omics datasets of large patient cohorts, dozens of best-practiced bioinformatics pipelines, and several AI technologies. Currently, the database includes the multi-omics data of more >20,000 cancer patients from about 40 cancer types, as well as about 2,000 cell lines and drug sensitivity data from >20,000 compounds. The main functional modules in this platform include (i) Target Discovery, (ii) Biomarker Identification, (iii) Drug Combination Prediction, and (iv) Clinical Trial Design. AIBERT focuses on clinically relevant signals with consistent patterns across different genotype-to-phenotype correlations and well balances the tradeoff between accuracy and robustness of data mining. This platform can effectively prioritize the candidate short lists for therapeutic targets, biomarkers, drug combinations or clinical investigations with high confidence and mechanistic insights.
Conclusions
An AI-driven, big data analytic platform, AIBERT provides a creative solution to address the major challenges in drug development, facilitating the paradigm shift from “hypothesis-driven” to big “data-driven”.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Precision Scientific (Beijing) Co., Ltd.
Funding
Has not received any funding.
Disclosure
X. Peng, Z. Li, H. Li, X. Guo, T. Sun: Full/Part-time employment: Precision Scientific (Beijing) Co., Ltd. X. Ji: Shareholder/Stockholder/Stock options: Precision Scientific (Beijing) Co., Ltd.