# Interpretable Deep Learning in Drug Discovery

## TL;DR

This paper explores methods to interpret neural networks in drug discovery, revealing how they identify pharmacophores and toxicophores, thus providing valuable insights for molecule design and understanding model predictions.

## Contribution

It introduces approaches to interpret neural network representations in drug discovery, including identifying relevant molecular features and extracting pharmacophores aligned with literature.

## Key findings

- Single neurons can classify pharmacophore-like structures
- Identified relevant molecular components for predictions
- Extracted pharmacophores are consistent with literature

## Abstract

Without any means of interpretation, neural networks that predict molecular properties and bioactivities are merely black boxes. We will unravel these black boxes and will demonstrate approaches to understand the learned representations which are hidden inside these models. We show how single neurons can be interpreted as classifiers which determine the presence or absence of pharmacophore- or toxicophore-like structures, thereby generating new insights and relevant knowledge for chemistry, pharmacology and biochemistry. We further discuss how these novel pharmacophores/toxicophores can be determined from the network by identifying the most relevant components of a compound for the prediction of the network. Additionally, we propose a method which can be used to extract new pharmacophores from a model and will show that these extracted structures are consistent with literature findings. We envision that having access to such interpretable knowledge is a crucial aid in the development and design of new pharmaceutically active molecules, and helps to investigate and understand failures and successes of current methods.