A team of researchers from King Abdullah University for Science and Technology claims to use a new deep learning approach that is considered “creative and combined” to find models and correlations between genes and diseases. The new algorithm uses different statistical models to identify any associations between the data.
In the same press release there is an example according to which by entering a sufficient number of tagged images as “Jack,” the system can then independently find other images with Jack without further suggestions.
However, the amount of data to be included in the programs in order to obtain automated learning is so high that the scientists themselves claim to have used a “creative” approach, so creative that Panagiotis Kalnis, a specialist in database management and information, and Xin Gao, a conventional life scientist, have developed an in-depth learning model that “goes beyond the current advanced methods.”
They combined several known data sets and were able to teach algorithms on how to identify diseases with similar manifestations, all using a deep learning model called a convolutional graphic network.
The result? The algorithm is able to identify complex and non-linear associations between genes and diseases, something that researchers can use to predict new associations. However, the same researchers still want to improve the precision quality of the software by giving it even more data.
Latest posts by Janice Walker (see all)
- Scientists discover that abundant minerals are proton conductors - November 2, 2019
- Scientists sequester the duckweed genome and discover genes that protect them from parasites - October 22, 2019
- Artificial intelligence learns complex patterns of genetic diseases - October 12, 2019