Imagine trying to solve a puzzle with one billion pieces. Imagine the feeling of accomplishment when you finish it, and then imagine how frustrating that same task would be if someone else completed all but two pieces for you! Immunologists face this dilemma every day as they try to predict T cell receptors’ specificity—the protein on their surface which helps them spot potential threats like viruses or bacteria.
IBM researchers are using deep learning to decode exactly what T cells do. They recently published a paper called TITAN: T Cell Receptor Specificity Prediction with Bimodal Neural Networks, where they explain in detail how IBM’s AI is able to predict the likelihood of any specific epitope binding to any specific receptor and vice versa.
The IBM Research team reveals that their AI outperforms the state-of-the-art method and provides biologically relevant explanations for its decisions. Their model could lead to using T cells as biomarkers to spot specific infections or cancers early, leading to more effective treatment options in a shorter time period than before.
Solving such a puzzle with machine learning is not trivial—there’s not enough data on receptors binding to epitopes. IBM researchers have only very few examples of puzzle pieces fitting together, so they decided to try a two-step approach drawing inspiration from previous work. They realized that predicting T cell receptor specificity is somewhat similar: The efficacy of the small molecule depends on its ability to bind and activate large protein molecules in targeted diseased tissue cells.
The IBM research team was able to use a technique called transfer learning in order for TITAN (T cell receptor-epitope interactions) to learn general concepts of chemical interactions from large datasets. They then fine-tuned the model by training it more precisely on T cell receptor-epitope interactions, which significantly boosted performance.
For the first time, IBM has solved a puzzle of sorts. The team is currently working on their long-term goal to build an algorithm that can reliably predict T cell receptor specificity, which could help make immunotherapies safer and better target cancer cells.
The Titan Project may change healthcare as we know it today – giving hope where there was none before and presenting new opportunities every day.
TITAN: T-cell receptor specificity prediction with bimodal attention networks [Paper]
Data-driven molecular design for discovery and synthesis of novel ligands: a case study on SARS-CoV-2 [Paper]
Current challenges for unseen-epitope TCR interaction prediction and a new perspective derived from image classification [Paper]
A message from Asif Razzaq, Co-founder of Marktechpost:
If you are a company looking to promote your product/webinar/conference/service, feel free to reach out via email to email@example.com. We offer sponsored posts and advertisements.
Asif Razzaq is an AI Journalist and Cofounder of Marktechpost, LLC. He is a visionary, entrepreneur and engineer who aspires to use the power of Artificial Intelligence for good.
Asif's latest venture is the development of an Artificial Intelligence Media Platform (Marktechpost) that will revolutionize how people can find relevant news related to Artificial Intelligence, Data Science and Machine Learning.
Asif was featured by Onalytica in it’s ‘Who’s Who in AI? (Influential Voices & Brands)’ as one of the 'Influential Journalists in AI' (https://onalytica.com/wp-content/uploads/2021/09/Whos-Who-In-AI.pdf). His interview was also featured by Onalytica (https://onalytica.com/blog/posts/interview-with-asif-razzaq/).