AI Can Now Predict How Cells React To Diseases And Treatments

Artificial intelligence is considered to be one of the greatest ‘big things’ of the century. But while the public actual understanding of AI is fairly basic and mostly through movies and fictions, AI is making leaps and bounds to become more like the futuristic, sci-fi portrayal in movies. Recently, researchers have developed an AI-assisted tool that has the promise to transform the way we do medical experimentation. Called scGen, medical researchers can use this tool to understand diseases and their treatments down to the tiny, cellular level.

"Generative Deep Learning Model"

The study and its results were published in the scientific journal Nature Methods. The article demoed how scGen can be a powerful tool that is capable of mapping the response of cells to diseases and the treatment administered beyond the realm of available experimental data.

ScGen, by nature, is a generative deep learning model that can utilize ideas derived from image, sequence, and language processing. These ideas will then be applied to help simulate cellular response through computer simulation.

With the proof of concept and prototype of ScGen working nominally, the only task left for the team is to improve and polish the tool. Making it completely data-driven and improve the algorithm to fortify its predictive power to expand its ability to study combinations of perturbations.

Alex Wolf - one of the researchers involved in the project coming from the Technical University of Munich, Germany commented that they can now start optimizing ScGena and fine-tuning it so that it would be able to answer more and more complex questions and handle a greater amount of data.

A Great Achievement in Computational Biology

A complete, large scale atlas of the healthy human organs will soon be made available. Specifically, through what is known as the Human Cell Atlas. This can be an extremely valuable tool for doctors. When cells, tissues, and organs in their healthy states are understood, the results can act as a reference for doctors to improve diagnosis and improve the efficiency in treating and monitoring diseases.

It has been a major goal in computational biology for quite some time: To successfully and accurately model cellular response to perturbations such as diseases, chemical, and biological actors e.t.c.

Out of Sample Capability

Furthermore, to mark the uniqueness of ScGen, it is the first tool to be able to predict the response of cells out of samples. If ScGen is trained to the effect of perturbations of one system, it can then derive and predict the response of another, completely unrelated system.

To put this into an even better perspective, Mohammad Lotfollahi - also one of the spearhead researchers from the Technical University of Munich stated that for the very first time, data generated from a mouse can predict response to disease and therapy in human.