DeepMind’s new technology can make predictions about genetic diseases
Bioinformatician Avsec: Predicting Rare Diseases by Missense Genetic Variants: How Good Will It Be in the Next 10 Years?
About 10 years ago, Žiga Avsec was a PhD physics student who found himself taking a crash course in genomics via a university module on machine learning. He was working in a lab that was studying rare diseases, and trying to figure out the exact genetic flaw that caused an unusual disease.
There are many genetic changes that cause a condition such as the one responsible for cystic fibrosis or the one that causes Sickle-cell disease. But researchers have observed only a few million of these single-letter ‘missense mutations’. Most of the 70 million possible in the human genome have no ill effect on health and only a small fraction have been linked to disease.
The missense variant that has been found in 4 million humans is labeled either benign or pathogenic after years of research. It can take months to study the effect of a single missense variant.
Computational predictions currently have a minimal role in diagnosing genetic diseases, says Marsh, and recommendations from physicians’ groups say that these tools should provide only supporting evidence in linking a mutation to a disease. Av Sec thinks that AlphaMissense has a bigger proportion of missense than have any previous methods. People will trust these models as they get better.
AlphaFold ushered in a new era in computational biology, but its impact won’t be as significant as this one. It is exciting. It is the best predictor at the moment. Will it be the most accurate predictor over the next two or three years? There’s a good chance it won’t be.”
Before applying AlphaMissense to thereal world, it must be rigorously evaluated with good performance metrics, as emphasized by the bioinformatician at Emory University.
A method called the Critical Assessment of Genome Interpretation (CAGI), is used to benchmark the performance of prediction methods for years against experimental data that has not been released. “It’s my worst nightmare to think of a doctor taking a prediction and running with it, as if it’s a real thing, without evaluation by entities such as CAGI,” Bromberg adds.
Bioinformatician iga Avsec has predicted rare diseases by using’AlphaMissense’ model of missense genetic variation. It can predict rare diseases by looking for genetic alterations that are linked to disease. The missense variant, found in over four million humans, is labelled either benign or pathogenic after years of research. The model claims to be the most accurate predictor for the next few years.