Filter feature selectors in the development of binary QSAR models

Examining evolutionary scale modeling-derived different-dimensional embeddings in the antimicrobial peptide classification through a KNIME workflow

Molecular features play an important role in different bio-chem-informatics tasks, such as the Quantitative Structure-Activity Relationships (QSAR) modeling. Several pre-trained models have been recently created to be used in downstream tasks, either by fine-tuning a specific model or by extracting features to feed traditional classifiers. In this regard, a new family of Evolutionary Scale Modeling models (termed as ESM-2 models) was recently introduced, demonstrating outstanding results in protein structure prediction benchmarks. Herein, we studied the usefulness of the different-dimensional embeddings derived from the ESM-2 models to classify antimicrobial peptides (AMPs). To this end, we built a KNIME workflow to use the same modeling methodology across experiments in order to guarantee fair analyses. As a result, the 640- and 1280-dimensional embeddings derived from the 30- and 33-layer ESM-2 models, respectively, are the most valuable  since statistically better performances were achieved by the QSAR models built from them. We also fused features of the different ESM-2 models, and it was concluded that the fusion contributes to getting better QSAR models than using features of a single ESM-2 model. Frequency studies revealed that only a portion of the ESM-2 embeddings is valuable for modeling tasks since between 43% and 66% of the features were never used. Comparisons regarding state-of-the-art deep learning (DL) models confirm that when performing methodologically principled studies in the prediction of AMPs, non-DL based QSAR models yield comparable-to-superior performances to DL-based QSAR models. The developed KNIME workflow is available-freely at https://github.com/cicese-biocom/classification-QSAR-bioKom. This workflow can be valuable to avoid unfair comparisons regarding new computational methods, as well as to propose new non-DL based QSAR models.

De Novo Generation and Identification of Novel Compounds with Drug Efficacy Based on Machine Learning

One of the main challenges in small molecule drug discovery is finding novel chemical compounds with desirable activity. Traditional drug development typically begins with target selection, but the correlation between targets and disease remains to be further investigated, and drugs designed based on targets may not always have the desired drug efficacy. The emergence of machine learning provides a powerful tool to overcome the challenge. Herein, a machine learning-based strategy is developed for de novo generation of novel compounds with drug efficacy termed DTLS (Deep Transfer Learning-based Strategy) by using dataset of disease-direct-related activity as input. DTLS is applied in two kinds of disease: colorectal cancer (CRC) and Alzheimer's disease (AD). In each case, novel compound is discovered and identified in in vitro and in vivo disease models. Their mechanism of actionis further explored. The experimental results reveal that DTLS can not only realize the generation and identification of novel compounds with drug efficacy but also has the advantage of identifying compounds by focusing on protein targets to facilitate the mechanism study. This work highlights the significant impact of machine learning on the design of novel compounds with drug efficacy, which provides a powerful new approach to drug discovery.

Machine learning analyses identify multi-modal frailty factors that selectively discriminate four cohorts in the Alzheimer's disease spectrum: a COMPASS-ND study

BACKGROUND

Frailty indicators can operate in dynamic amalgamations of disease conditions, clinical symptoms, biomarkers, medical signals, cognitive characteristics, and even health beliefs and practices. This study is the first to evaluate which, among these multiple frailty-related indicators, are important and differential predictors of clinical cohorts that represent progression along an Alzheimer's disease (AD) spectrum. We applied machine-learning technology to such indicators in order to identify the leading predictors of three AD spectrum cohorts; viz., subjective cognitive impairment (SCI), mild cognitive impairment (MCI), and AD. The common benchmark was a cohort of cognitively unimpaired (CU) older adults.

METHODS

The four cohorts were from the cross-sectional Comprehensive Assessment of Neurodegeneration and Dementia dataset. We used random forest analysis (Python 3.7) to simultaneously test the relative importance of 83 multi-modal frailty indicators in discriminating the cohorts. We performed an explainable artificial intelligence method (Tree Shapley Additive exPlanation values) for deep interpretation of prediction effects.

RESULTS

We observed strong concurrent prediction results, with clusters varying across cohorts. The SCI model demonstrated excellent prediction accuracy (AUC = 0.89). Three leading predictors were poorer quality of life ([QoL]; memory), abnormal lymphocyte count, and abnormal neutrophil count. The MCI model demonstrated a similarly high AUC (0.88). Five leading predictors were poorer QoL (memory, leisure), male sex, abnormal lymphocyte count, and poorer self-rated eyesight. The AD model demonstrated outstanding prediction accuracy (AUC = 0.98). Ten leading predictors were poorer QoL (memory), reduced olfaction, male sex, increased dependence in activities of daily living (n = 6), and poorer visual contrast.

CONCLUSIONS

Both convergent and cohort-specific frailty factors discriminated the AD spectrum cohorts. Convergence was observed as all cohorts were marked by lower quality of life (memory), supporting recent research and clinical attention to subjective experiences of memory aging and their potentially broad ramifications. Diversity was displayed in that, of the 14 leading predictors extracted across models, 11 were selectively sensitive to one cohort. A morbidity intensity trend was indicated by an increasing number and diversity of predictors corresponding to clinical severity, especially in AD. Knowledge of differential deficit predictors across AD clinical cohorts may promote precision interventions.

Graph-Based Feature Selection Approach for Molecular Activity Prediction

In the construction of QSAR models for the prediction of molecular activity, feature selection is a common task aimed at improving the results and understanding of the problem. The selection of features allows elimination of irrelevant and redundant features, reduces the effect of dimensionality problems, and improves the generalization and interpretability of the models. In many feature selection applications, such as those based on ensembles of feature selectors, it is necessary to combine different selection processes. In this work, we evaluate the application of a new feature selection approach to the prediction of molecular activity, based on the construction of an undirected graph to combine base feature selectors. The experimental results demonstrate the efficiency of the graph-based method in terms of the classification performance, reduction, and redundancy compared to the standard voting method. The graph-based method can be extended to different feature selection algorithms and applied to other cheminformatics problems.

Explainable Molecular Sets: Using Information Theory to Generate Meaningful Descriptions of Groups of Molecules

Algorithmically identifying the meaningful similarities between an assortment of molecules is a critical chemical problem, and one which is only gaining in relevance as data-driven chemistry continues to progress. Effectively addressing this challenge can be achieved through a reformulation of the problem into information theory, cluster-based supervised classification, and the implementation of key concepts, particularly information entropy and mutual information. These concepts are combined with unsupervised learning atop learned chemical spaces to generate meaningful labels for arbitrary collections of molecules. An open-source and highly extensible codebase is provided to undertake these experiments, demonstrate the viability of the approach on known clusters, and glean insights into the learned representations of chemical space within message-passing neural networks, an architecture not readily permitting interpretability. This approach facilitates the interoperability between human chemical knowledge and the algorithmically derived insights, which will continue to become more prevalent in the coming years.

QSAR model for predicting neuraminidase inhibitors of influenza A viruses (H1N1) based on adaptive grasshopper optimization algorithm

High-dimensionality is one of the major problems which affect the quality of the quantitative structure-activity relationship (QSAR) modelling. Obtaining a reliable QSAR model with few descriptors is an essential procedure in chemometrics. The binary grasshopper optimization algorithm (BGOA) is a new meta-heuristic optimization algorithm, which has been used successfully to perform feature selection. In this paper, four new transfer functions were adapted to improve the exploration and exploitation capability of the BGOA in QSAR modelling of influenza A viruses (H1N1). The QSAR model with these new quadratic transfer functions was internally and externally validated based on MSE_train, Y-randomization test, MSE_test, and the applicability domain (AD). The validation results indicate that the model is robust and not due to chance correlation. In addition, the results indicate that the descriptor selection and prediction performance of the QSAR model for training dataset outperform the other S-shaped and V-shaped transfer functions. QSAR model using quadratic transfer function shows the lowest MSE_train. For the test dataset, proposed QSAR model shows lower value of MSE_test compared with the other methods, indicating its higher predictive ability. In conclusion, the results reveal that the proposed QSAR model is an efficient approach for modelling high-dimensional QSAR models and it is useful for the estimation of IC₅₀ values of neuraminidase inhibitors that have not been experimentally tested.

The Influence of Structural Patterns on Acute Aquatic Toxicity of Organic Compounds

Investigation of the influence of molecular structure of different organic compounds on acute toxicity towards Fathead minnow, Daphnia magna, and Tetrahymena pyriformis has been carried out using 2D simplex representation of molecular structure and two modelling methods: Random Forest (RF) and Gradient Boosting Machine (GBM). Suitable QSAR (Quantitative Structure - Activity Relationships) models were obtained. The study was focused on QSAR models interpretation. The aim of the study was to develop a set of structural fragments that simultaneously consistently increase toxicity toward Fathead minnow, Daphnia magna, Tetrahymena pyriformis. The interpretation allowed to gain more details about known toxicophores and to propose new fragments. The results obtained made it possible to rank the contributions of molecular fragments to various types of toxicity to aquatic organisms. This information can be used for molecular optimization of chemicals. According to the results of structural interpretation, the most significant common mechanisms of the toxic effect of organic compounds on Fathead minnow, Daphnia magna and Tetrahymena pyriformis are reactions of nucleophilic substitution and inhibition of oxidative phosphorylation in mitochondria. In addition acetylcholinesterase and voltage-gated ion channel of Fathead minnow and Daphnia magna are important targets for toxicants. The on-line version of the OCHEM expert system (https://ochem.eu) were used for a comparative QSAR investigation. The proposed QSAR models comply with the OECD principles and can be used to reliably predict acute toxicity of organic compounds towards Fathead minnow, Daphnia magna and Tetrahymena pyriformis with allowance for applicability domain estimation.

Predictive Modeling for Frailty Conditions in Elderly People: Machine Learning Approaches

Background

Frailty is one of the most critical age-related conditions in older adults. It is often recognized as a syndrome of physiological decline in late life, characterized by a marked vulnerability to adverse health outcomes. A clear operational definition of frailty, however, has not been agreed so far. There is a wide range of studies on the detection of frailty and their association with mortality. Several of these studies have focused on the possible risk factors associated with frailty in the elderly population while predicting who will be at increased risk of frailty is still overlooked in clinical settings.

Objective

The objective of our study was to develop predictive models for frailty conditions in older people using different machine learning methods based on a database of clinical characteristics and socioeconomic factors.

Methods

An administrative health database containing 1,095,612 elderly people aged 65 or older with 58 input variables and 6 output variables was used. We first identify and define six problems/outputs as surrogates of frailty. We then resolve the imbalanced nature of the data through resampling process and a comparative study between the different machine learning (ML) algorithms – Artificial neural network (ANN), Genetic programming (GP), Support vector machines (SVM), Random Forest (RF), Logistic regression (LR) and Decision tree (DT) – was carried out. The performance of each model was evaluated using a separate unseen dataset.

Results

Predicting mortality outcome has shown higher performance with ANN (TPR 0.81, TNR 0.76, accuracy 0.78, F1-score 0.79) and SVM (TPR 0.77, TNR 0.80, accuracy 0.79, F1-score 0.78) than predicting the other outcomes. On average, over the six problems, the DT classifier has shown the lowest accuracy, while other models (GP, LR, RF, ANN, and SVM) performed better. All models have shown lower accuracy in predicting an event of an emergency admission with red code than predicting fracture and disability. In predicting urgent hospitalization, only SVM achieved better performance (TPR 0.75, TNR 0.77, accuracy 0.73, F1-score 0.76) with the 10-fold cross validation compared with other models in all evaluation metrics.

Conclusions

We developed machine learning models for predicting frailty conditions (mortality, urgent hospitalization, disability, fracture, and emergency admission). The results show that the prediction performance of machine learning models significantly varies from problem to problem in terms of different evaluation metrics. Through further improvement, the model that performs better can be used as a base for developing decision-support tools to improve early identification and prediction of frail older adults.

Improving (Q)SAR predictions by examining bias in the selection of compounds for experimental testing

Existing data on structures and biological activities are limited and distributed unevenly across distinct molecular targets and chemical compounds. The question arises if these data represent an unbiased sample of the general population of chemical-biological interactions. To answer this question, we analyzed ChEMBL data for 87,583 molecules tested against 919 protein targets using supervised and unsupervised approaches. Hierarchical clustering of the Murcko frameworks generated using Chemistry Development Toolkit showed that the available data form a big diffuse cloud without apparent structure. In contrast hereto, PASS-based classifiers allowed prediction whether the compound had been tested against the particular molecular target, despite whether it was active or not. Thus, one may conclude that the selection of chemical compounds for testing against specific targets is biased, probably due to the influence of prior knowledge. We assessed the possibility to improve (Q)SAR predictions using this fact: PASS prediction of the interaction with the particular target for compounds predicted as tested against the target has significantly higher accuracy than for those predicted as untested (average ROC AUC are about 0.87 and 0.75, respectively). Thus, considering the existing bias in the data of the training set may increase the performance of virtual screening.