How can SHAP values help to shape metabolic stability of chemical compounds?

Abstract

Background Computational methods support nowadays each stage of drug design campaigns. They assist not only in the process of identification of new active compounds towards particular biological target, but also help in the evaluation and optimization of their physicochemical and pharmacokinetic properties. Such features are not less important in terms of the possible turn of a compound into a future drug than its desired affinity profile towards considered proteins. In the study, we focus on metabolic stability, which determines the time that the compound can act in the organism and play its role as a drug. Due to great complexity of xenobiotic transformation pathways in the living organisms, evaluation and optimization of metabolic stability remains a big challenge. Results Here, we present a novel methodology for the evaluation and analysis of structural features influencing metabolic stability. To this end, we use a well-established explainability method called SHAP. We built several predictive models and analyse their predictions with the SHAP values to reveal how particular compound substructures influence the model’s prediction. The method can be widely applied by users thanks to the web service, which accompanies the article. It allows a detailed analysis of SHAP values obtained for compounds from the ChEMBL database, as well as their determination and analysis for any compound submitted by a user. Moreover, the service enables manual analysis of the possible structural modifications via the provision of analogous analysis for the most similar compound from the ChEMBL dataset. Conclusions To our knowledge, this is the first attempt to employ SHAP to reveal which substructural features are utilized by machine learning models when evaluating compound metabolic stability. The accompanying web service for metabolic stability evaluation can be of great help for medicinal chemists. Its significant usefulness is related not only to the possibility of assessing compound stability, but also to the provision of information about substructures influencing this parameter. It can assist in the design of new ligands with improved metabolic stability, helping in the detection of privileged and unfavourable chemical moieties during stability optimization. The tool is available at https://metstab-shap.matinf.uj.edu.pl/.

An application based on bioinformatics and machine learning for risk prediction of sepsis at first clinical presentation using transcriptomic data

“…Lundberg et al developed the SHAP model, which uses the SHAP value as a uniform measure of the importance of features used in the machine learning models ( Lundberg et al, 2020 ). By attributing output values to the Shapley value of each feature, researchers have performed interpretability analysis of machine learning models ( Wang et al, 2021 ; Wojtuch et al, 2021 ; Scavuzzo et al, 2022 ). In this study, high gene expression of TP6V1D had a positive impact on prediction, whereas low gene expression of ATP6V1D negatively impacted prediction, similar to CLIC1 .…”

Explaining structure-activity relationships using locally faithful surrogate models

“…[50] Rodríguez-Pérez and Bajorath [51] showed how SHAP can be used to generate local explanations for compound activity predictions. Wojtuch et al [40] used SHAP to understand metabolic activity of compounds using Molecules are described using Molecular ACCess System (MACCS) fingerprints. Figure 1d is an example of feature importances extracted using SHAP.…”

“…Counterfactuals, for example, are actionable and contextual since they provide exact changes that need to be made to a molecule to change its activity. They are agnostic to input [35] (b) Contrastive explanations identify the missing features that may influence the prediction, image taken from Lim et al [37] (c)Atomic attribution techniques give scores for each contribution of atoms and subgroups [33,39] (d) SHAP uses Shapley values to give feature attributions [40] (e) Gradient based methods for graph attribution [41] features. Feature attribution and weighting methods are limited by the original set of features or model inputs.…”

“…Figure 1: Different explanation methods from literature. (a) Counterfactuals give the smallest possible change that changes the activity[35] (b) Contrastive explanations identify the missing features that may influence the prediction, image taken from Lim et al[37] (c)Atomic attribution techniques give scores for each contribution of atoms and subgroups[33,39] (d) SHAP uses Shapley values to give feature attributions[40] (e) Gradient based methods for graph attribution[41] …”

A triple-classification for differentiating renal oncocytoma from renal cell carcinoma subtypes and CK7 expression evaluation: a radiomics analysis

“…However, previous studies lacked the interpretability of radiomics models, which led to skepticism about the underlying mechanisms of the radiomics features. In the current study, we explained our classifiers by Shapley additive explanations (SHAP) framework to increase their usability [ 23 ]. Currently, SHAP is the most recommended tool for model explanation.…”