State-of-the-art AI programs can support drug development by predicting the interaction of proteins with small molecules. However, researchers at the University of Basel show that these programs only memorize patterns instead of understanding physical relationships. They often fail when it comes to new proteins that would be particularly interesting for innovative drugs.
An American-Korean research team from Dallas, Seattle, Seoul, and Wonju has developed a deep learning model for predicting protein complex structures in order to predict what proteins interact with which other proteins. Around 190 million human protein pairs were analyzed. They succeeded in predicting around 5,500 previously unknown interactions between human proteins, including those involved in immunity, metabolism, and cell signaling.
Researchers at Helmholtz Zentrum München as well as the Technical University of Munich (TUM) have developed a so-called NicheFormer. It is the first large-scale foundation model that combines single-cell analysis with spatial transcriptomics.
An international challenge invites creative minds to explore how artificial intelligence can strengthen and further develop the Adverse Outcome Pathway Framework.
This month, Prof. Dr. Alexander Mosig took up the newly established professorship for alternative methods to animal testing in infection and inflammation research at Jena University Hospital.
An Italian research team from Genoa used a MIVO® esophagus-on-chip platform to test the protective and regenerative effects of two different active ingredients for protecting the mucous membrane. Their test results showed that GSE Reflusolve Rapid achieved a better effect in the model compared to the second test substance, which was an alginate- and carbonate-based drug.
At the end of September, the National Institutes of Health (NIH) announced the contract award for the establishment of a Standardized Organoid Modeling (SOM) Center. Using artificial intelligence, robotics, and a variety of human cell sources, the new NIH SOM Center will standardize organoid models and make them accessible and reproducible.
A research team has established an in vitro menstrual cycle (IVMC) protocol using human endometrial organoids that accurately replicates the epithelium during the key phases of the menstrual cycle, including differentiation, hormone withdrawal, degradation, and regeneration.
Researchers at Utrecht University Medical Center have expanded their in vivo research on mice with animal-free methods after receiving support from the 3Rs Centre Utrecht. Suppo3Rt is a staffed central support desk that offers tailored guidance to researchers and other stakeholders about the possibilities of NAMs as a replacement or reduction of animal testing.
The startup LifeTaq from Klosterneuburg, Austria, in collaboration with the Austrian Research Institute for Chemistry and Technology (OFI), has developed a robot-based machine that is capable of producing 3D tissue models fully automatically under controlled conditions. It is intended to replace animal testing in preclinical research.