Kode Chemoinformatics launches ALChemy, the suite for Automated Learning for Chemistry
Thanks to the integration of its new product QUEEN to FAST, the recently launched features selection tool, ALChemy simplifies the whole QSAR modelling process.
September 21st 2023, Pisa – Kode Chemoinformatics, the Business Unit of Kode focused on chemoinformatics and AI projects in the chemistry, pharma, food and biotechnology fields, launches ALChemy (Automated Learning for Chemistry), the suite taking care of the whole QSAR/QSPR (Quantitative Structure–Activity/Property Relationship) modelling process, from the feature selection to the training of the predictive model.
This suite is the result of the integration of two products, developed by the Chemoinformatics BU of Kode, which can be executed either separately or in pipeline:
FAST – Feature Analysis and Selection Tool
Launched a few months ago, FAST is an application designed to automate the selection of molecular descriptors for QSAR/QSPR modelling.
Since none of the many available techniques for feature selection presents theoretical advantages over the others (nor a criterion to establish a priori the best approach exists), the idea behind this product is to systematically search and test hundreds of different solutions obtained by as many methods.
This allows to evaluate and select the solution that shows the best a posteriori performance, by exploiting a sequential approach that, reducing step by step the dataset size, makes it manageable even on a standard laptop.
QUEEN – QSAR Under Effective and Efficient Neural-Networks
In the past few years, QSAR models powered by neural networks have gained popularity due to their ability to capture intricate and nonlinear relationships in chemical data. Although they often require a complex parametrization, these models have emerged as the most powerful and versatile algorithms for handling this type of problem.
Queen is designed to break down this complexity.
- a simple and intuitive import method for training data
- a fully automated hyperparameters optimization algorithm based on Bayesian method
- a model training section
- the possibility to merge hundreds of individual models into a single response.
The tool is completed by an export function that allows the final model saving. The exported file can be then transferred, reloaded, and used to predict the target property on any new sample. Each prediction comes with an applicability domain score that indicates the reliability of the model outcome for the given sample.
The exported model can be reimported and run both on QUEEN and QUEEN Runner, the free licence version of QUEEN made for running existing models only.
ALChemy: the suite for Automated Learning for Chemistry
Just two months after the launch of FAST, Kode Chemoinformatics successfully finalised the development of QUEEN, thus expanding the suite’s capacity to handle the full range of tasks involved in the QSAR modelling process.
ALChemy represents a powerful support for industries and research institutions in the invaluable job of identifying potential candidates with desired properties.
The efficient development and release of a comprehensive suite in such a brief timeframe was made possible by ideals and mission of Kode: ensuring unparalleled responsiveness in casting solutions that save time, resources, and effort in domains ranging from environmental science to pharmaceutical and ecotoxicological research.
Alessio Sommovigo, Kode Chemoinformatics BU Manager, declares: “We are proud of this suite. There are few solutions in the chemoinformatics market that set such a high goal: supporting the entire modelling process. With ALChemy, we want to make it easier for all computational chemists to profit from cutting-edge data science technologies, by guaranteeing robust and performing results under all conditions, while keeping a simple and almost parameter-free setup”.
Marco Calderisi, CEO of Kode, adds: “It is a pleasure to contribute to the broadening of the use of statistics in chemistry. In fact, at Kode we see every day how multidisciplinary teams with different skills can lead to excellent results. Indeed, we firmly believe that contamination (of skills, first and foremost, but not only) is the way to solve many of the challenges we face on a daily basis. The ALChemy Suite is proof of this”.
