We used AI Design Labs to design a Neural Network based molecule fingerprinting to encode both biological as well as structural information. Traditional algorithms usually utilize only structural information, and hence are less reliable for biological activity prediction without processing. We used our neural network to fingerprint DrugBank's small molecule database to study drug activity. Learn more and view an open access demo of Neural Fingerprinting here.
We have developed an automated deep learning development pipeline to help you easily create in silico prediction algorithms for real-world datasets (such as experimental observations, wet-lab data). We provide support for converting raw data into a feature-rich format for deep learning. Our deep learning pipeline is flexible and can be configured to work even in situations with low data availability. Contact us to learn more.
We developed ChemFaiss as an open-source solution for providing a way to perform an AI-based search for chemical data (Find out more here). It utilizes chemical fingerprinting to search for similar molecules within large databases and uses parallelization to speed up the process. Our neural fingerprinter in combination with ChemFaiss allows us to look for existing drugs with similar bioactivity. This enables us to study properties of new molecules using existing data, search for alternatives, and discover novel uses for existing drugs. Contact us to learn more.
Bring deep learning to your drug development process with AIDDT AI labs. We make it easy to build and securely train AI models on your dataset.
We provide support for automatically building and refining features from your data. Our software automatically converts a variety of chemical and bioinformatics data into a format on which deep learning models can be trained.
We received your message and will contact you back soon.
Error sending please try again