ARKA

One of the most commonly used in silico approaches for toxicity assessment of small datasets is the Quantitative Structure-Activity Relationship (QSAR), which generates predictive models for the efficient prediction of query compounds. However, the reliability of the predictions from QSARs derived from small datasets is often questionable from a statistical point of view. This is due to the presence of a larger number of descriptors as compared to the number of training compounds, which reduces the degree of freedom of the developed model. To reduce the overall prediction error for a particular QSAR model, we have proposed the computation of the novel Arithmetic Residuals in K-groups Analysis (ARKA) descriptors. We have reduced the number of modeling descriptors in a supervised manner by partitioning them into K classes (K = 2 here) depending on the higher mean normalized values of the descriptors to a particular response class, thus preventing the loss of chemical information. A scatter plot of the data points using the values of two ARKA descriptors (ARKA_2 vs. ARKA_1) can potentially identify activity cliffs, less confident data points, and less modelable data points. We have used [1]  five representative environmentally relevant endpoints (skin sensitization, earthworm toxicity, milk/plasma partitioning, algal toxicity, and rodent carcinogenicity of hazardous chemicals) with graded responses to which the ARKA framework was applied for classification modeling.  On comparing the performance of the models generated using conventional QSAR descriptors and the ARKA descriptors, the prediction quality of the models derived from ARKA descriptors was found, based on multiple graded-data validation metrics-derived decision criteria, much better than the models derived from QSAR descriptors signifying the potential of ARKA descriptors in ecotoxicological classification modeling of small data sets. Additionally, this holds true for the Read-Across approach as well, since the Read-Across predictions using ARKA descriptors supersede the predictions generated from QSAR descriptors. This approach can not only be used in assessing environmental/ecotoxicity endpoints but can also be extended to other fields like drug discovery.

Reference: [1] Banerjee A, Roy K, ARKA: A framework of dimensionality reduction for machine-learning classification modeling, risk assessment, and data gap-filling of sparse environmental toxicity data. Environ Sci: Process Impacts, 2024, https://doi.org/10.1039/D4EM00173G 

Disclaimer: The figures on this page are taken from the above publication.

Java-based Tool

ARKA 

Tool developed by Arkaprava Banerjee

License agreement form and Google form for usage

MultiClass ARKA

Publications Reporting ARKA Descriptors from Other Laboratories 

• Classification and Regression Machine Learning Models for Predicting the Combined Toxicity and Interactions of Antibiotics and Fungicides Mixtures. Environmental Pollution (2024).

• Evaluating ionic liquid toxicity with machine learning and structural similarity methods. Green Chemical Engg. (2024)

• Development of a robust Machine learning model for Ames test outcome prediction. Chem Phys Lett (2024)

• Comparative QSAR and q-RASAR Modeling for Aquatic Toxicity of Organic Chemicals to Three Trout Species: O. Clarkii, S. Namaycush, and S. Fontinalis. J Hazard Mater (2024)

• New binary mixtures of fungicides against Macrophomina phaseolina: machine learning-driven QSAR, read-across prediction, and molecular dynamics simulation. Chemosphere (2024)

• Contributions to the development of prediction models for the toxicity of ionic liquids. Struct Chem (2024)

• Explainable machine learning models for predicting the acute toxicity of pesticides to sheepshead minnow (Cyprinodon variegatus). Sci Tot Environ (2024)

• Unveiling the interspecies correlation and sensitivity factor analysis of rat and mouse acute oral toxicity of antimicrobial agents: first QSTR and QTTR Modeling report. Toxicology Research, 13, tfae191 (2024)

• A comprehensive machine learning-based models for predicting mixture toxicity of azole fungicides toward algae (Auxenochlorella pyrenoidosa). Environ Internat (2024)

• Development of Quantitative Structure Property Relationship Models and Tool for Predicting the Soil Adsorption Coefficient (logKOC). Environ Pollut (2025)

• Web Server-based Deep Learning-Driven Predictive Models for Respiratory Toxicity of Environmental Chemicals: Mechanistic Insights and Interpretability. J Hazard Mater (2025)

• PBScreen: A Server for the High-Throughput Screening of Placental Barrier–Permeable Contaminants Based on Multifusion Deep Learning. Environ Pollut (2025)

• Explainable machine learning models enhance prediction of PFAS bioactivity using quantitative molecular surface analysis-derived representation. Water Research (2025)

• Modeling and Interpretability Study of the Structure–Activity Relationship for Multigeneration EGFR Inhibitors. ACS Omega (2025)

• Risk Assessment of Industrial Chemicals Towards Salmon Species Amalgamating QSAR, q-RASAR, and ARKA Framework. Toxicol Reports (2025)

• Predicting chemical toxicity towards Raphidocelis subcapitata with quantum chemical descriptors. Algal Research (2025) 104055

• Evaluating the Vascular Risk of PFCs: An Integrated XGBoost-Driven Structure−Activity Prediction and Experimental Validation Study. Environ & Health (2025)

• Development of an Interpretable Machine Learning Model for Neurotoxicity Prediction of Environmentally Related Compounds. Environ Sci Tech (2025)

• QSAR Model Development for the Environmental Risk Limits and High-Risk List Identification of Phenylurea Herbicides in Aquatic Environments. J Agric Food Chem (2025)

• Estimation of Activity Coefficient of Aqueous Ionic Liquids Using a Machine Learning method: The Artificial Neural Network coupled with Group Contribution Approach. J Indian Chem Soc (2025)

• Structure-activity relationships of halogenated disinfectants as potent inhibitors of 5α-reductase type 1: Potential impact on neurosteroid synthesis. J Hazard Mater 496, 139498 (2025)

Publications on the Multi-class ARKA approach from other laboratories

• QSAR-Based Prediction of Acute Inhalation Toxicity and SHAP Interpretability Analysis of Fluorocarbon Environmental-friendly Insulating Gases. Environ Res (2025)

• Unlocking Structural Insights into CO2 Absorption with Ionic Liquids: A Comparative Study of Machine Learning, Association Rules, and Meta-Learning for Modeling Henry’s Law Constant. ACS Sustainable Chemistry & Engineering (2025)

Publications on ARKA descriptors from the DTC Lab in collaboration with other laboratories

• Organic Sunscreens and Their Products of Degradation in Biotic and Abiotic Conditions—In Silico Studies of Drug-Likeness and Human Placental Transport. Int. J. Mol. Sci. 25(22), 12373 (2024)

Publications on ARKA descriptors from the DTC Lab 

• Development of Machine Learning-Based Models for Mutagenicity Predictions with Applications to Non-Sugar Sweeteners. Mol Informatics 44 (2025) e2400357

• A New Approach Methodology (NAM) for carcinogenicity prediction of organic chemicals using the Multiclass ARKA framework and machine-learning-based stacking regression. J Hazard Mater (2025)

ARKA Presentation

BCADD2024 - Kunal Roy