Expanding Predictive Capacities in Toxicology: Insights from Hackathon-Enhanced Data and Model Aggregation.

In the realm of predictive toxicology for small molecules, the applicability domain of QSAR models is often limited by the coverage of the chemical space in the training set. Consequently, classical models fail to provide reliable predictions for wide classes of molecules. However, the emergence of innovative data collection methods such as intensive hackathons have promise to quickly expand the available chemical space for model construction. Combined with algorithmic refinement methods, these tools can address the challenges of toxicity prediction, enhancing both the robustness and applicability of the corresponding models. This study aimed to investigate the roles of gradient boosting and strategic data aggregation in enhancing the predictivity ability of models for the toxicity of small organic molecules. We focused on evaluating the impact of incorporating fragment features and expanding the chemical space, facilitated by a comprehensive dataset procured in an open hackathon. We used gradient boosting techniques, accounting for critical features such as the structural fragments or functional groups often associated with manifestations of toxicity.

Electrochemical platform for detecting Escherichia coli bacteria using machine learning methods.

We present an electrochemical platform designed to reduce time of Escherichia coli bacteria detection from 24 to 48-h to 30 min. The presented approach is based on a system which includes gallium-indium (eGaIn) alloy to provide conductivity and a hydrogel system to preserve bacteria and their metabolic species during the analysis. The work is dedicated to accurate and fast detection of Escherichia coli bacteria in different environments with the supply of machine learning methods. Electrochemical data obtained during the analysis is processed via multilayer perceptron model to identify i.e. predict bacterial concentration in the samples. The performed approach provides the effectiveness of bacteria identification in the range of 102-109 colony forming units per ml with the average accuracy of 97%. The proposed bioelectrochemical system combined with machine learning model is prospective for food analysis, agriculture, biomedicine.

Electrochemical Sensor to Detect Antibiotics in Milk Based on Machine Learning Algorithms.

The present study is dedicated to the problem of electrochemical analysis of multicomponent mixtures, such as milk. A combination of cyclic voltammetry facilities and machine learning techniques made it possible to create a pattern recognition system for the detection of antibiotic residues in skimmed milk. A multielectrode sensor including copper, nickel, and carbon fiber was fabricated for the collection of electrochemical data. Processes occurring at the electrode surface were discussed and simulated with the help of molecular docking and density functional theory modeling. It was assumed that the antibiotic fingerprint reveals a potential drift of electrodes, owing to complexation with metal ions present in milk. The gradient boosting algorithm showed the best efficiency in training the machine learning model. High accuracy was achieved for the recognition of antibiotics in milk. The elaborated method may be incorporated into existing milking systems at dairy farms for monitoring the residue concentrations of antibiotics.

When Bubbles Are Not Spherical: Artificial Intelligence Analysis of Ultrasonic Cavitation Bubbles in Solutions of Varying Concentrations.

Ultrasonic irradiation of liquids, such as water-alcohol solutions, results in cavitation or the formation of small bubbles. Cavitation bubbles are generated in real solutions without the use of optical traps making our system as close to real conditions as possible. Under the action of the ultrasound, bubbles can grow, oscillate, and eventually collapse or decompose. We apply the mathematical method of separation of motions to interpret the acoustic effect on the bubbles. While in most situations, the spherical shape of a bubble is the most energetically profitable as it minimizes the surface energy, when the acoustic frequency is in resonance with the natural frequency of the bubble, shapes with the dihedral symmetry emerge. Some of these resonance shapes turn unstable, so the bubble decomposes. It turns out that bubbles in the solutions of different concentrations (with different surface energies and densities) attain different evolution paths. While it is difficult to obtain a deterministic description of how the solution concentration affects bubble dynamics, it is possible to separate images with different concentrations by applying the artificial neural network (ANN) algorithm. An ANN was trained to detect the concentration of alcohol in a water solution based on the bubble images. This indicates that artificial intelligence (AI) methods can complement deterministic analysis in nonequilibrium, near-unstable situations.