In this issue:
Welcome back to your weekly dose of AI news for Life Science!
This week, we have some exciting new models lined up for you:
Dive into these game-changing innovations and explore how they are transforming the biotech and healthcare landscapes!
RhoDesign: RNA Aptamer design 🧬
RNA molecules are essential biomolecules, acting as cellular "translators" and carrying out tasks from storing and transmitting genetic information to catalyzing biochemical reactions. Among these, RNA aptamers are a unique class of sequences designed to bind specific target molecules. Aptamers hold promise across various fields, including diagnostics, biosensing, and therapeutic development. However, designing new, functional RNA aptamers, particularly those with specific structural requirements, remains challenging. Introducing RhoDesign, a sequence-to-structure model that generates RNA aptamers based on target three-dimensional shapes!
📌 Key Insights:
RhoDesign was trained on ~370,000 RhoFold-predicted 3D RNA structures representative of different types of RNA and ~3,500 3D RNA structures from the Protein Data Bank
The authors tested RhoDesign in designing better aptamers against the TO1–biotin small molecule. The top 20 aptamers designed against were synthesised and tested, with 15 of the 20 aptamers were active, and 9 of the 15 active aptamers exhibited higher activity compared to a baseline aptamer
Overall RhoDesign outperformed existing models across a variety of metrics, demonstrating its capability to generate RNA structures that closely mimic target aptamers and maintain functional accuracy.
BALM - Protein-Ligand binding affinity prediction 🪢
Accurate in-silico prediction of protein-ligand binding affinity is essential for efficient hit identification in large molecular libraries. Commonly used structure-based methods such as giga-docking often fail to rank compounds effectively, and free energy-based approaches, while accurate, are too computationally intensive for large-scale screening. Existing deep learning models struggle to generalize to new targets or drugs, and current evaluation methods do not reflect real-world performance accurately. Introducing BALM, a deep learning framework that predicts binding affinity using pretrained protein and ligand language models!
📌 Key Insights:
BALM combines both a protein (ESM2) and a ligand (ChemBERTa-2) language model to strongly generalise to unseen drugs, scaffolds, and targets
BALM excels in both easy and difficult scenarios, showing better results compared to traditional tools like AutoDock and Vina
A small fine-tuning (ie. extra training) of <1% of the parameters resulted in an 18% improvement of binding prediction
CyanobacteriaDB : Cyanobacteria Bioactive Compounds Database for Drug Development💿
Cyanobacteria strains are increasingly being explored since we realized their potential to produce bioactive compounds that can be used in therapeutics and bioremediation. However, there is no clear space where data regarding such compounds are stored, or at least, can be easily extracted and used safely (using the FAIR principles). Introducing CyanoBioActiveDB, a comprehensive database that can be used for both therapeutic applications and environmental bioremediation. The authors demonstrate the ability of their ML model to properly identify potential protein targets of cyanobacterial compounds with unknown bioactivity information, and thus, it is a valuable tool for future molecular dynamics and docking studies!
📌 Key Insights:
A Python programming protocol with 3431 cyanobacteria bioactive compounds, 373 unique protein targets, and 3027 molecular descriptors
Updatable, Curated, Searchable, Downloadable, and follows the FAIR principles
The information stored in the CyanobioactiveDB database is usable in virtual screening and docking computational techniques
Did you find this newsletter insightful? Share it with a colleague!
Subscribe Now to stay at the forefront of AI in Life Science.
Connect With Us
Have questions or suggestions? We'd love to hear from you!
📧 Email Us | 📲 Follow on LinkedIn | 🌐 Visit Our Website