Selection of Informative Examples in Chemogenomic Datasets

The present state and challenges of active learning in drug discovery

Active learning (AL) is an iterative feedback process that efficiently identifies valuable data within vast chemical space, even with limited labeled data. This characteristic renders it a valuable approach to tackle the ongoing challenges faced in drug discovery, such as the ever-expanding explore space and the limitations of labeled data. Consequently, AL is increasingly gaining prominence in the field of drug development. In this paper, we comprehensively review the application of AL at all stages of drug discovery, including compounds-target interaction prediction, virtual screening, molecular generation and optimization, as well as molecular properties prediction. Additionally, we discuss the challenges and prospects associated with the current applications of AL in drug discovery.

Analysis of the fatigue status of medical security personnel during the closed-loop period using multiple machine learning methods: a case study of the Beijing 2022 Olympic Winter Games

Using machine learning methods to analyze the fatigue status of medical security personnel and the factors influencing fatigue (such as BMI, gender, and wearing protective clothing working hours), with the goal of identifying the key factors contributing to fatigue. By validating the predicted outcomes, actionable and practical recommendations can be offered to enhance fatigue status, such as reducing wearing protective clothing working hours. A questionnaire was designed to assess the fatigue status of medical security personnel during the closed-loop period, aiming to capture information on fatigue experienced during work and disease recovery. The collected data was then preprocessed and used to determine the structural parameters for each machine learning algorithm. To evaluate the prediction performance of different models, the mean relative error (MRE) and goodness of fit (R2) between the true and predicted values were calculated. Furthermore, the importance rankings of various parameters in relation to fatigue status were determined using the RF feature importance analysis method. The fatigue status of medical security personnel during the closed-loop period was analyzed using multiple machine learning methods. The prediction performance of these methods was ranked from highest to lowest as follows: Gradient Boosting Regression (GBM) > Random Forest (RF) > Adaptive Boosting (AdaBoost) > K-Nearest Neighbors (KNN) > Support Vector Regression (SVR). Among these algorithms, four out of the five achieved good prediction results, with the GBM method performing the best. The five most critical parameters influencing fatigue status were identified as working hours in protective clothing, a customized symptom and disease score (CSDS), physical exercise, body mass index (BMI), and age, all of which had importance scores exceeding 0.06. Notably, working hours in protective clothing obtained the highest importance score of 0.54, making it the most critical factor impacting fatigue status. Fatigue is a prevalent and pressing issue among medical security personnel operating in closed-loop environments. In our investigation, we observed that the GBM method exhibited superior predictive performance in determining the fatigue status of medical security personnel during the closed-loop period, surpassing other machine learning techniques. Notably, our analysis identified several critical factors influencing the fatigue status of medical security personnel, including the duration of working hours in protective clothing, CSDS, and engagement in physical exercise. These findings shed light on the multifaceted nature of fatigue among healthcare workers and emphasize the importance of considering various contributing factors. To effectively alleviate fatigue, prudent management of working hours for security personnel, along with minimizing the duration of wearing protective clothing, proves to be promising strategies. Furthermore, promoting regular physical exercise among medical security personnel can significantly impact fatigue reduction. Additionally, the exploration of medication interventions and the adoption of innovative protective clothing options present potential avenues for mitigating fatigue. The insights derived from this study offer valuable guidance to management personnel involved in organizing large-scale events, enabling them to make informed decisions and implement targeted interventions to address fatigue among medical security personnel. In our upcoming research, we will further expand the fatigue dataset while considering higher precisionprediction algorithms, such as XGBoost model, ensemble model, etc., and explore their potential contributions to our research.

DeepReac+: deep active learning for quantitative modeling of organic chemical reactions

Various computational methods have been developed for quantitative modeling of organic chemical reactions; however, the lack of universality as well as the requirement of large amounts of experimental data limit their broad applications. Here, we present DeepReac+, an efficient and universal computational framework for prediction of chemical reaction outcomes and identification of optimal reaction conditions based on deep active learning. Under this framework, DeepReac is designed as a graph-neural-network-based model, which directly takes 2D molecular structures as inputs and automatically adapts to different prediction tasks. In addition, carefully-designed active learning strategies are incorporated to substantially reduce the number of necessary experiments for model training. We demonstrate the universality and high efficiency of DeepReac+ by achieving the state-of-the-art results with a minimum of labeled data on three diverse chemical reaction datasets in several scenarios. Collectively, DeepReac+ has great potential and utility in the development of AI-aided chemical synthesis. DeepReac+ is freely accessible at https://github.com/bm2-lab/DeepReac.

Active learning effectively identifies a minimal set of maximally informative and asymptotically performant cytotoxic structure-activity patterns in NCI-60 cell lines

The NCI-60 cancer cell line screening panel has provided insights for development of subtype-specific chemical therapies and repurposing. By extracting chemical structure and cytotoxicity patterns, virtual screening potentially complements the availability of high-throughput assay platforms and improves bioactive compound discovery rates by computational prefiltering of candidate compound libraries. Many groups report high prediction performances in computational models of NCI-60 data when using cross-validation or similar techniques, yet prospective therapy development in novel cancers may have little to no such data and further may not have the resources to perform hit identification using large compound libraries. In contrast to bulk screening and analysis, the active learning methodology has demonstrated how to identify compounds for screening in small batches and update computational models iteratively, leading to predictive models with a minimum number of compounds, and importantly clarifying data volumes at which limits in predictive ability are achieved. Here, in replicate per-cell line experiments using 50% of data (∼20 000 compounds) as the external prediction target, predictive limits are reproducibly demonstrated at the stage of systematic selection of 10-30% of the incorporable half. The pattern was consistent across all 60 cell lines. Limits of predictability are found to be correlated to the doubling times of cell lines and the number of cellular response discontinuities (activity cliffs) present per cell line. Organization into chemical scaffolds delineated degrees of predictive challenge. These results provide key insights for strategies in developing new inhibitors in existing cell lines or for future automated therapy selection in personalized oncotherapy.

Practical considerations for active machine learning in drug discovery

Active machine learning enables the automated selection of the most valuable next experiments to improve predictive modelling and hasten active retrieval in drug discovery. Although a long established theoretical concept and introduced to drug discovery approximately 15 years ago, the deployment of active learning technology in the discovery pipelines across academia and industry remains slow. With the recent re-discovered enthusiasm for artificial intelligence as well as improved flexibility of laboratory automation, active learning is expected to surge and become a key technology for molecular optimizations. This review recapitulates key findings from previous active learning studies to highlight the challenges and opportunities of applying adaptive machine learning to drug discovery. Specifically, considerations regarding implementation, infrastructural integration, and expected benefits are discussed. By focusing on these practical aspects of active learning, this review aims at providing insights for scientists planning to implement active learning workflows in their discovery pipelines.

Applicability Domain of Active Learning in Chemical Probe Identification: Convergence in Learning from Non-Specific Compounds and Decision Rule Clarification

Efficient identification of chemical probes for the manipulation and understanding of biological systems demands specificity for target proteins. Computational means to optimize candidate compound selection for experimental selectivity evaluation are being sought. The active learning virtual screening method has demonstrated the ability to efficiently converge on predictive models with reduced datasets, though its applicability domain to probe identification has yet to be determined. In this article, we challenge active learning’s ability to predict inhibitory bioactivity profiles of selective compounds when learning from chemogenomic features found in non-selective ligand-target pairs. Comparison of controls versus multiple molecule representations de-convolutes factors contributing to predictive capability. Experiments using the matrix metalloproteinase family demonstrate maximum probe bioactivity prediction achieved from only approximately 20% of non-probe bioactivity; this data volume is consistent with prior chemogenomic active learning studies despite the increased difficulty from chemical biology experimental settings used here. Feature weight analyses are combined with a custom visualization to unambiguously detail how active learning arrives at classification decisions, yielding clarified expectations for chemogenomic modeling. The results influence tactical decisions for computational probe design and discovery.