Researchers have long been searching for new medicines, but the process of trawling through massive collections of chemical compounds to find those that could become lifesaving therapies is tedious and time-consuming. Now, a team of scientists has made breakthrough using artificial intelligence (AI) to speed up this process.
According to Hongkui Deng, a cell biologist who was not involved in the work, the new AI method is "a powerful blueprint for the future" which creates a 'smart' screening system that learns from its own experiments. This could be game-changer in drug discovery as it can evaluate how compounds perturb entire networks of gene activity.
In traditional drug discovery methods, researchers would test tens of thousands of compounds on cells grown in laboratory to identify those with potential therapeutic effects. However, this method is expensive and laborious, especially when integrating large screens with complex assays.
To find a tractable way to harness genomic data, Alex Shalek, a biomedical engineer at the Massachusetts Institute of Technology, teamed up with other researchers and Cellarity, a biotechnology company. Together, they trained a deep-learning model called DrugReflector on publicly available data about how each chemical compound affects gene activity in more than 50 kinds of cells.
The team used DrugReflector to find chemicals that can affect the generation of platelets and red blood cells - a characteristic that could be useful in treating some blood conditions. They then tested 107 of these chemicals to determine whether they had the predicted effect. Overall, the team found that DrugReflector was up to 17 times more effective at finding relevant compounds than traditional drug screening methods.
Moreover, when the researchers incorporated the data from their first round of screening into the model, its success rate doubled. This breakthrough could potentially turbocharge the hunt for new medicines by harnessing the power of AI and genomic data.
According to Hongkui Deng, a cell biologist who was not involved in the work, the new AI method is "a powerful blueprint for the future" which creates a 'smart' screening system that learns from its own experiments. This could be game-changer in drug discovery as it can evaluate how compounds perturb entire networks of gene activity.
In traditional drug discovery methods, researchers would test tens of thousands of compounds on cells grown in laboratory to identify those with potential therapeutic effects. However, this method is expensive and laborious, especially when integrating large screens with complex assays.
To find a tractable way to harness genomic data, Alex Shalek, a biomedical engineer at the Massachusetts Institute of Technology, teamed up with other researchers and Cellarity, a biotechnology company. Together, they trained a deep-learning model called DrugReflector on publicly available data about how each chemical compound affects gene activity in more than 50 kinds of cells.
The team used DrugReflector to find chemicals that can affect the generation of platelets and red blood cells - a characteristic that could be useful in treating some blood conditions. They then tested 107 of these chemicals to determine whether they had the predicted effect. Overall, the team found that DrugReflector was up to 17 times more effective at finding relevant compounds than traditional drug screening methods.
Moreover, when the researchers incorporated the data from their first round of screening into the model, its success rate doubled. This breakthrough could potentially turbocharge the hunt for new medicines by harnessing the power of AI and genomic data.
, i've been saying it for ages that platforms like cellality are gonna change the game in biotech research! no more tedious sifting through chemical compounds, this AI method is genius 
, 17 times more effective than traditional methods? that's crazy! 
and now they can incorporate data from previous rounds of screening, so it's only a matter of time before we see some groundbreaking new medicines 
. as a platform loyalist, i'm all about embracing innovation like this 
i'm so hyped for this breakthrough and the potential it holds for saving lives 
. the fact that they used publicly available data to train their model is genius 
& its success rate of up to 17 times more effective is mind-blowing 
! i'm all about harnessing the power of AI in med discovery 
, can't wait to see what other breakthroughs come out of this tech 
Also, 17 times more effective than traditional methods? That sounds too good to be true 
. I'm all for innovation, but let's not get ahead of ourselves just yet 
. This 'smart' screening system learns from its own experiments and can even evaluate how compounds affect entire networks of gene activity 
. But now, with AI on the team, they're finding relevant compounds 17 times faster than before 
. It's like science fiction come true! 
. This could be a game-changer for finding new medicines, especially when it comes to treating blood conditions. But what do you guys think about the potential risks of relying too heavily on AI in this field? Should we be worried about losing human intuition in the lab? 
.
.
. The fact that it can evaluate how compounds affect entire networks of gene activity is wild 
. But let's be real, we've seen similar "breakthroughs" before 
? I'm not saying it won't work, but we need to see some more concrete results before getting too hyped 
.
It's like, who needs to manually go through tons of chemical compounds when you can have a super smart computer do it for you? 
The fact that they were able to use publicly available data to train their model is genius 
And the best part is, it's already 17x more effective than traditional methods 
. Can't wait to see what else AI comes up with in the future 
. I mean, who needs to test tens of thousands of compounds on cells when you have an AI that can learn from its own experiments and find the right ones? It's a game-changer, trust me 
. I remember when I was in college, we used to do all this manual work with spreadsheets and stuff, it took forever! This AI method is like having a superpower 
! And with AI and genomic data, they can learn from their own experiments and improve the model over time.