NOV 09, 2019 04:41 AM PST

Using machine learning to discover new cancer treatments

Flashback fifty years ago and many of the modern world’s inventions and discoveries would have been undreamed of – including using machines to discover new anti-cancer drugs. But that’s exactly what’s happening. New research published in eLife by scientists at Sanford Burnham Prebys Medical Discovery Institute details how an algorithm is helping to develop epigenetic drug screens and the potential that this technology has for cancer and other diseases.

Scientists used machine learning to discover new anti-cancer developments. Photo: Pixabay

Senior author Alexey Terskikh, Ph.D. is an associate professor in Sanford Burnham Prebys' Development, Aging and Regeneration Program. Terskikh commented on the study saying, "In order to identify the rare few drug candidates that induce desired epigenetic effects, scientists need methods to screen hundreds of thousands of potential compounds. Our study describes a powerful image-based approach that enables high-throughput epigenetic drug discovery."

The image-based approach Terskikh is referring to is Microscopic Imaging of Epigenetic Landscapes (MIEL), a type of high-content phenotypic screening that is more advanced than past technologies in its ability to, as the authors write, “capture the nuclear staining patterns of epigenetic marks and employ machine learning to accurately distinguish between such patterns.”

Is all that sounding a little complicated? Let’s break it down in simpler terms by first grasping the concept of epigenetics. Watch the video below to understand how epigenetics works.

A little clearer now? So, what the researchers in Terskikh’s lab did was take advantage of a machine-learning algorithm to track the epigenetic changes in cells for more than 220 drugs, in order to determine just how the compounds are working. Through this process, they were able to identify certain drugs that can target and potentially treat glioblastoma, the most fatal kind of brain cancer. They say this method could be useful in identifying compounds that could treat other types of cancer, as well as heart disease and mental illness.

The researchers are hoping to put their algorithms into the commercial field as soon as possible. "Our method is ready for immediate use by pharmaceutical companies looking to develop epigenetic drug screens," says first author Chen Farhy, Ph.D., a postdoctoral researcher in the Terskikh lab. "Industry and academic researchers working on mechanistic studies may also benefit from this method, as the algorithm can detect and categorize epigenetic changes induced by experimental treatments, genetic manipulations or other approaches."

Sources: eLife, Science Daily

About the Author
  • Kathryn is a curious world-traveller interested in the intersection between nature, culture, history, and people. She has worked for environmental education non-profits and is a Spanish/English interpreter.
You May Also Like
NOV 21, 2019
Cancer
NOV 21, 2019
Could electromagnetic fields treat metastatic triple-negative breast cancer?
New research published in Communications Biology suggests that electromagnetic fields are capable of stopping the metastasis of some breast cancer cells. I...
NOV 21, 2019
Cancer
NOV 21, 2019
The role of sleeper cells in metastasis
New research published in the journal Nature Communications aims to explain why some breast cancers recur following cancer treatment. According to col...
NOV 21, 2019
Cancer
NOV 21, 2019
Bias against unmarried cancer patients
A new study published in The New England Journal of Medicine examines the biases behind the cancer treatments that doctors prescribe, coming to the conclus...
NOV 21, 2019
Genetics & Genomics
NOV 21, 2019
A Way to Predict Which Microbes Can Cause Cancer
Researchers have created a technique that can identify bacteria and viruses that are linked to cancer....
NOV 21, 2019
Cancer
NOV 21, 2019
How HPV might actually defend against skin cancer
New research from scientists at Massachusetts General Hospital features findings that suggest that immunity to certain strains of HPV (human papillomavirus...
NOV 21, 2019
Cancer
NOV 21, 2019
When the numbers don't match: inaccuracy in late state breast cancer survival estimates
According to the World Health Organization, about 2.1 million women are diagnosed with breast cancer annually. It is the cancer that most affects women and...
Loading Comments...