stub AI Researchers Develop Method to Repurpose Existing Drugs to Fight Covid-19 - Unite.AI
Connect with us

Artificial Intelligence

AI Researchers Develop Method to Repurpose Existing Drugs to Fight Covid-19

Updated on

An international team of researchers has applied AI models to find already existing drugs that can treat Covid-19 in elderly patients. The research team applied autoencoder models to drugs already on the market, aiming to find commonalities in changes to gene expression caused by both natural aging and Covid-19.

As explained by study co-author computational biologist at MIT, Caroline Uhler, the problem with developing new drugs to combat Covid-19 is that the drug development process can take years. AI has already been leveraged to discover new drugs, finding new formulations for therapeutic drugs much faster than traditional drug discovery methods. Unfortunately, even the relatively fast speed at which drugs can be discovered using AI is still far too slow to when it comes to situations like the Covid-19 pandemic. It’s much more expedient to repurpose existing drugs.

In order to find a drug that can fight off the effects of Covid-19 in elderly populations, the researchers looked at genes that underwent changes during both normal aging and when affected by the Covid-19 virus.

Covid-19 is hypothesized to use certain cell pathways, particularly inflammatory pathways, to replicate. It’s also known that the effects of Covid-19 are much worse in elderly populations than younger populations. Furthermore, the respiratory systems of aging individuals are characterized by alterations in tissue stiffness. Given these facts, the researchers looked for genes altered by both aging and Covid-19, with the goal being to find drugs that interact positively with these genes.

The research team used a three-step process to find genes common to both pathways. In the first phase of the research, the team used an autoencoder to generate a list of candidate drugs. This was done by having the autoencoder analyze two datasets of gene expression patterns, selecting the drugs that appeared to reduce the virus’ overall impact. The result was a list of candidate drugs and their accompanying interactions with proteins in both aging and infection pathways. Afterward, the researchers took the list of candidate drugs and mapped the interactions between proteins and the two different pathways, producing a map of protein interactions for both. The researchers then compared the two protein interaction maps to find areas of overlap. This lead to the discovery of a gene expression network that drugs should target to reduce the severity of Covid-19 in older patients.

In the final phase of the research project, the team employed statistical methods to determine causality within the mapped networks. Using this method, they were able to determine the exact genes that a drug candidate should interact with in order to most effectively reduce the severity of a Covid-19 infection.

According to the results of their analysis, the RIPK1 gene was the part of the genome thought to be most suited for targeting by Covid-19 therapy drugs. Some of the candidate drugs are used to treat cancer. Other candidate drugs are already being tested by medical institutes to treat Covid-19.

The research team notes that this is just the first step in determining which drugs could be repurposed for the treatment of Covid-19. Extensive in vitro experiments and clinical trials will have to be carried out to determine if the drugs are actually effective. However, if the approach proves successful it could be used to find effective drugs for other conditions.

According to the researcher team writes:

“While we apply our computational platform in the context of SARS-CoV-2, our algorithms integrate data modalities that are available for many diseases, thereby making them broadly applicable.”