-
Categories
-
Pharmaceutical Intermediates
-
Active Pharmaceutical Ingredients
-
Food Additives
- Industrial Coatings
- Agrochemicals
- Dyes and Pigments
- Surfactant
- Flavors and Fragrances
- Chemical Reagents
- Catalyst and Auxiliary
- Natural Products
- Inorganic Chemistry
-
Organic Chemistry
-
Biochemical Engineering
- Analytical Chemistry
- Cosmetic Ingredient
-
Pharmaceutical Intermediates
Promotion
ECHEMI Mall
Wholesale
Weekly Price
Exhibition
News
-
Trade Service
"Alpha Fold 2" realizes open source |
Will help develop more powerful protein structure prediction technology |
It means that scientists will widely use software that can accurately determine the 3D shape of proteins
.
On July 16, the London-based company DeepMind released an open source version of its deep learning neural network AlphaFold 2, and described its method in a paper in the journal Nature
Understanding the structure of a protein helps determine the function of the protein and understand the role of various mutations
.
Up to now, the structure of about 100,000 proteins has been analyzed experimentally, but this accounts for only a small part of the billions of proteins that have been sequenced
John Jumper, Demis Hassabis and colleagues of DeepMind in London, UK, described AlphaFold2, which is a new model based on neural networks that predicts protein structure with atomic level accuracy.
The author claims that the method has reached "Unprecedented accuracy"
.
From May to July last year, Jumper et al.
validated this method in the 14th "Critical Evaluation of Protein Structure Prediction" (CASP14) competition
.
The competition requires the participating teams to analyze the structure of proteins based on their amino acid sequences
In the competition, most of the structures predicted by AlphaFold2 reached unprecedented accuracy, not only comparable to experimental methods, but also far superior to other methods for analyzing new protein structures
.
The protein structure obtained by the experimental method is superimposed on the structure of AlphaFold2, and the median distance (95% coverage) between the superimposed atoms constituting the backbone of the protein backbone is 0.
AlphaFold2's neural network can predict the structure of a typical protein within a few minutes, as well as the structure of a larger protein (such as a protein containing 2180 amino acids and no homologous structure)
.
The model can accurately predict the reliability of its prediction based on each amino acid, which is convenient for researchers to use its prediction results
The author believes that this precise prediction algorithm can allow protein structure analysis technology to keep up with the pace of development of the genome revolution
.
Demis Hassabis, founder and CEO of DeepMind, said that after the company unveiled a new AlphaFold system that accurately predicts the 3D structure of proteins to the atomic level at the CASP14 conference last year, it promised to share its methods and provide the scientific community with extensive and free access.
Way
.
"Today we took the first step of our commitment
.
" Hassabis said, "We look forward to seeing what other new methods it will inspire, and we look forward to sharing more of our new developments with you soon
The structure of the binding of human interleukin-12 protein to its receptor predicted by machine learning software
.
Image credit: Ian Haydon, Institute of Protein Design Medicine, University of Washington
Links to related articles:
https://doi.