Mapping the avoid-ome: a systematic open-science approach to predictive ADMET.

Area of Science:

• Drug discovery and development
• Computational chemistry
• Structural biology

Background:

• Drug discovery frequently encounters setbacks (30% of clinical failures) due to unpredictable Absorption, Distribution, Metabolism, Excretion, and Toxicity (ADMET) properties.
• Current methods lack the detailed, atomistic insights required to address the
• The
• Avoid-ome
• - a set of proteins acting as
• anti-targets
• - presents a significant challenge in rational drug design.

Purpose of the Study:

• To establish OpenADMET, an open-science initiative focused on generating pre-competitive, mechanistic datasets.
• To address the limitations of conventional drug discovery methods by providing atomistic detail on
• anti-targets
• .

Main Methods:

• Utilizing high-throughput structural biology techniques.
• Employing active learning strategies to refine models.
• Leveraging community challenges to foster collaboration and data generation.

Main Results:

• Development of generalizable predictive models grounded in structural
• ground truth
• .

Conclusions:

• OpenADMET directly studies the
• Avoid-ome
• to overcome ADMET-related failures.
• Facilitating a new era of rational, multi-parameter drug design through open science and mechanistic datasets.