Artificial Intelligence Finds Mutations Faster
August 20, 2012

Artificial Intelligence Used To Examine Mutant Worms

Michael Harper for — Your Universe Online

In order to fully understand some diseases – how they work, how they spread, response to drugs, etc – researchers have been employing some of the world´s tiniest animals, such as the multicellular fruit fly, nematode, or zebra fish. While this examination by itself is useful, these researchers can make real progress in their studies if they can examine a multitude of animals in order to detect any mutation among them.

Now, scientists have been able to create such a system which will allow a mass observation as the researchers try to pinpoint mutations in these animals. With these mutations observed, researchers will be able to continue their disease studies on one kind of animal, free of mutation.

This new automated system created by Georgia Tech scientists uses Artificial Intelligence (AI) and brand new, cutting-edge image processing to accurately and quickly study large groups of a specific kind of nematode used in biological research, Caenorhabditis elegans (c. elegans) to be exact.

This new, automated system not only replaces the old, manual method of determining differences in each worm, it is also able to detect even subtle differences between worms, making it much more accurate than the previous methods. In fact, this new, AI-driven method can detect mutations which would have previously gone undetected.

The scientists are understandably excited about this new method, saying the ability to quickly scan thousands of worms in a fraction of the time could dramatically change the way they genetically screen the C. elegans nematodes.

Details of this research appeared in an advance, online publication of the journal Nature Methods on August 19th.

“While humans are very good at pattern recognition, computers are much better than humans at detecting subtle differences, such as small changes in the location of dots or slight variations in the brightness of an image,”  said the project´s lead researcher Hang Lu. Lu is an associate professor in the School of Chemical & Biomolecular Engineering at the Georgia Institute of Technology.

“This technique found differences that would have been almost impossible to pick out by hand.”

Specifically, Lu´s team is looking for genes which affect the way synapses are formed and developed inside the worms. With this research, Lu and his team hope they´ll be able to transfer their findings into the understanding of human brain development. To carry out this study, the researchers create mutations in the genomes in thousands of worms. Then, the researchers closely watch the worms to detect any changes in their synapses. When a worm carrying a mutation is detected, it is removed for further analysis and study to better understand why these synapses changed.

Sometimes, these synapses are formed in the wrong place in the worm´s genetic makeup, while other times they are the wrong size or wrong type of synapse. Lu´s team of researchers are studying these mutations specifically to understand any developmental patterns in the mutant worms.

As these researchers examine the worms for these mutations, the new AI-driven machine takes 3D images of each worm as it passes through the sorter. The machine then compares each worm to the “wild type,” or genetically mutated worms, that the team has already taught the machine to be on the lookout for.

“We feed the program wild-type images, and it teaches itself to recognize what differentiates the wild type. It uses this information to determine what a mutant type may look like which is information we didn´t provide to the system and sorts the worms based on that,” said Matthew Crane, a graduate student who helped carry out the research.

“We don´t have to show the computer every possible mutant, and that is very powerful. And the computer never gets bored.”

“We are hoping that the technology will really change the approach people can take to this kind of research,” said a hopeful Lu.

“We expect that this approach will enable people to do much larger scale experiments that can push the science forward beyond looking what individual mutations are doing in a specific situation.”