Friday, February, 15, 2019 11:58:08

Google’s AI specialist DeepMind has hit the headlines for what is being considered a significant milestone regarding the use of AI in predicting 3D structures of proteins depending entirely on their genetic sequence.

In terms of disease diagnostics and treatment, it is essential to understand the protein structures, as they can improve the scientists’ understanding of the human body, while possibly supporting protein design and bioengineering.

As per trusted sources, the 3D models of proteins generated by AlphaFold are way more precise as compared to any that have come before, thereby leading to significant progress in a core challenge of biology.

According to sources equipped with the knowledge of the matter, there are multiple methods of predicting the native 3D state of protein molecules from residual amino acids in DNA. However, modeling the 3D structure is an intricate task, considering how many permutations can exist due to protein folding relying on factors like interactions between amino acids, cite sources.

Reportedly, there even exists a game called FoldIt, that tries to leverage human intuitions to estimate workable protein forms. According to DeepMind, its approach is based on years of prior research associated with the use of big data to predict the protein structure. Furthermore, it is applying deep learning approaches exclusively to genomic data.

Fortunately, the field of genomics has sufficient data due to the rapid reduction in the costs of genetic sequencing. Therefore, deep learning methods to predict problems that are dependent on genomic data has gained significant popularity over the last few years. DeepMind’s work on this problem led to AlphaFold, which it submitted to CASP this year, cite trusted sources.

Although the technology needs a lot more work before it can have a measurable impact on treating diseases, dealing with the environment, and a lot more, the company is aware of its enormous potential, states a source familiar with the matter.

With a dedicated team focused on how machine learning can help science, the company is anticipating the many ways its technology can make a difference, claim reports.