The new model predicts the production of MRNA protein, the acceleration of the drug and the vaccine

A new model of artificial intelligence can improve the process of discovering drugs and vaccines by predicting how effective mRNA sequences will produce protein, both in general and in different types of cells. The new progress, developed through an academic-industrial cooperation between the University of Texas in Austin and Sanofi, contributes to the prediction of how many protein cells will be able to minimize the need to experiment with testing and error, accelerating the next generation of therapeutic mRNA.

RNA Messenger (MRNA) contains instructions for which proteins need to do and how to do it, allowing our bodies to grow and perform the daily processes of life. Among the most promising areas of health and medicine, the ability to develop new MRNA vaccines and drugs – capable of combating viruses, cancers and genetic disorders – includes the often provocative process of patient cells to produce enough protein from therapeutic mRNA.

The new model, called Ribonn, is to guide the design of the new mRNA -based therapy by illuminating what will deliver the highest amount of a protein or better specific parts of the body such as the heart or liver. The team described their model today in one of the two relevant documents in Nature Biotechnology.

“When we started this project six years ago, there was no obvious application. We were curious if the cells coordinate which MRNAS produce and how effectively they translate into proteins, this is the value of curiosity -based research.

Can Cenik, Associate Professor, Molecular Biosciences, University of Texas, Austin

Work was made possible with the funding of support by the National Institutes of Health, the Welch Foundation and the Lonestar6 supercomputer at the UT Texas Advanced Computing Center.

In tests that extend to more than 140 types of human and mouse cells, Ribonn was about twice the prediction of translation efficiency such as previous approaches. This progress can lend researchers to make predictions to cells in ways that could help accelerate cancer treatments and infectious and hereditary diseases.

You can think about how the cells in your body make protein as the way a chef can bake. To cook a batch of protein, the chefs in one of your cells (ribosomes) are looking for the recipe in your own unique cookbook (Aka DNA), copy the recipe to notebooks called RNAS (MRNAS) messengers and then combine the components.

A mRNA vaccine or therapeutically evenly these chefs in your cells for protein production. In the case of a vaccine, they could produce a protein found on the surface of a pathogenic virus or cancer cells, essentially waving a large red flag in front of your immune system to create antibodies to the virus or cancer. In the case of a disorder caused by a genetic mutation, they may produce a protein that your body cannot do properly by reversing the disorder.

Prior to the development of their new predictive model, Cenik and the UT team were first edited a set of public data available from over 10,000 experiments that measure how effectively different mRNA translates into proteins into different types of human cells and mice. Once this had created this training set, AI experts and mechanical learning by UT and Sanofi met to develop Ribonn.

A goal of the prediction tool is one day to do treatments aimed at a particular type of cells, said Cenik, who is also a subsidiary school at the ODEN Institute for UT computing engineering and sciences and a CPRIT scholar, who receives support from the Institute of Cancer Institute and Cancer Institute.

“You may need a next -generation treatment that needs to be done in the liver or lung or immune system cells,” he said. “This opens the opportunity to change the mRNA sequence to increase the production of this protein in this cell type.”

In a accompanying document also in nature biotechnology, the group has shown that MRNAs with related biological functions translate into proteins at similar levels into different cell types. Scientists have long been aware that the process of transcription of genes with related functions to MRNAS is coordinated, but it had not been proven that mRNA translation into proteins is also coordinated.