Friday, July 30

Create the most complete map of human proteins using AI | Digital Trends Spanish

DeepMind, a research laboratory for Artificial Intelligence (AI), created the most comprehensive map of human proteins to date using AI.

The firm is a subsidiary of Alphabet, Google’s parent company, and some scientists already compare the potential impact of this work with the Human Genome Project.


Proteins are long, complex molecules that carry out various tasks in the body, from building tissues to fighting disease.

According to the researchers, its purpose is dictated by its structure, which folds like origami into complex and irregular shapes.


Understanding how it folds helps explain protein function, while helping scientists with various tasks, from fundamental research on how the body works to designing new drugs.

Until now, determining the structure of a protein was based on costly and time-consuming experiments, but last year DeepMind showed that it could make accurate predictions of a protein’s structure using AI software called AlphaFold.

“I see this as the culmination of DeepMind’s 10+ years of life,” the company’s CEO and co-founder, Demis Hassabis, told The Verge.

“From the beginning, this is what we set out to do: to advance AI, to test that in games like Go Y Atari, [y] apply that to real world problems to see if we can accelerate scientific advances and use them to benefit humanity. “

Deepmind publishes predictions of the structure of some 350,000 proteins from 20 different organisms, including animals such as mice, the fruit fly, and bacteria such as E. coli.

Most importantly, the publication includes predictions for 98 percent of all human proteins, some 20,000 different structures, which are known as the human proteome.

Although it is not the first public data set of this type of macromolecules, it is the most complete and accurate.

Editor’s Recommendations

var stage = decodeURIComponent(0); var options = JSON.parse(decodeURIComponent('')); var allOptions = {};

if (stage > 0 && window.DTOptions) { allOptions = window.DTOptions.getAll();

Object.keys(options).forEach(function(groupK) { if (options[groupK] && typeof options[groupK] === 'object') { Object.keys(options[groupK]).forEach(function(k) { if (!allOptions[groupK] || typeof allOptions[groupK] !== 'object') { allOptions[groupK] = {}; }

allOptions[groupK][k] = options[groupK][k]; }); } }); } else { allOptions = options; }

var getAll = function () { return allOptions; };

var get = function (key, group, def) { key = key || ''; group = group || decodeURIComponent('qnqb92BhrzmkpqGx'); def = (typeof def !== 'undefined') ? def : null;

if (typeof allOptions[group] !== 'undefined') { if (key && typeof allOptions[group][key] !== 'undefined') { return allOptions[group][key]; } }

return def; };

var set = function (key, group, data) { key = key || ''; group = group || decodeURIComponent('qnqb92BhrzmkpqGx'); data = data || null;

if (key) { if (typeof allOptions[group] === 'undefined') { allOptions[group] = {}; }

allOptions[group][key] = data; } };

var del = function (key, group) { key = key || ''; group = group || decodeURIComponent('qnqb92BhrzmkpqGx');

if (typeof allOptions[group] !== 'undefined') { if (key && typeof allOptions[group][key] !== 'undefined') { allOptions[group][key] = null; } } };

window.DTOptions = { get: get, getAll: getAll, set: set, del: del, }; }());

Leave a Reply

Your email address will not be published. Required fields are marked *