New AI Does Faster, Cheaper Gene Analysis To Help Cure Diseases Like Alzheimer's

Today is a memorable day in the field of genetic research as scientists from GPU maker Nvidia as well as Harvard have made a major breakthrough. They have created a deep learning toolkit that slashes the time cost it usually takes to run rare, single-cell experiments.

Getty Images

Reported first in Nature Communications, the toolkit dubbed AtacWorks is capable of running inference on an entire genome. In case you thought that this isn¡¯t a big deal, that process normally takes over two days easily, but with the toolkit in action, this time is brought down to just half an hour, which is just insane.

There are various elements responsible for this crazy speed, but the primary one is Nvidia¡¯s Tensor Core GPUs that boost AI-related processing. Moreover, AtacWorks functions in tandem with ATAC-seq -- a popular method that is actually fabricated to find out open areas in the genome of healthy and diseased cells.

These areas are actually small parts of a person¡¯s DNA which can be instrumental in determining and activating particular functions. These parts can be helpful to offer scientists early indication if a person could have Alzheimer¡¯s, heart disease or cancer.

Normally, ATAC-sec requires analysis of thousands of cells to get to a proper conclusion, however, AtacWorks can offer a similar result with just ten of such cells. What¡¯s more, is that they used AtacWorks with a dataset of stem cells that produced red and white blood cells.

These subtypes can¡¯t really be studied with the help of conventional methods, however, with AtacWorks they could identify different parts of the DNA linked with white and red blood cells, respectively.

How will this impact everything?

You might be wondering how all this AI mumbo-jumbo will help better our lives. Well, this crazy capability offers the ability to analyse a genome faster and in a more affordable manner.

iStock

This could be just the thing we need in order to identify specific mutations or biomarkers to lead us to the source of specific diseases (diseases like Alzheimer's, Dementia, Parkinson¡¯s etc.), diseases that are incurable and later also help us crack the drug code that could beat the disease right in the bud.

NVIDIA researcher and the paper¡¯s lead author Avantika Lal explains, ¡°With very rare cell types, it¡¯s not possible to study differences in their DNA using existing methods. AtacWorks can help not only drive down the cost of gathering chromatin accessibility data, but also open up new possibilities in drug discovery and diagnostics.¡±