Courtesy of AlphaFold
In December of 2020, I wrote about AlphaFold, the AI program from Alphabet (Google) that revolutionized picturing proteins. Proteins are structures of Amino acids that "fold" into strange shapes. Scientists had been working on protein structures laboriously until AlphaFold, an offshoot of DeepMind, started printing out structures by the hundreds. In that Musing, I wrote:
AI has lifted the lid off protein folding.
I know. What is protein folding? It's difficult, but stick around for what could be one of the headline-creating stories of this decade.
I was too conservative.
In July 2021, DeepMind and EMBL-EBI made the AlphaFold database public and freely accessible to users worldwide. By the beginning of 2022, the curated database had grown to contain one million protein structure predictions, and the ability to predict protein structures from amino acid sequences was coined scientific breakthrough of the year 2021.
They have finished the human proteome.
Partners use AlphaFold, the AI system recognised last year as a solution to the protein structure prediction problem, to release more than 350,000 protein structure predictions including the entire human proteome to the scientific community.
AlphaFold has finished the entire protein structure for humans and is starting to work on the proteins of lab animals (rats, pigs, monkeys). In 18 months AlphaFold and another AI program RoseTTAFold have revolutionized medical research. Not only can they predict the structure of human proteins, they can predict the shape of combined proteins and ones that have become misshapen usually indicating a problem like cancer.
Recent predictions of complex protein interactions can shed light on the mechanistic machinery underlying important biological processes. In October 2021, DeepMind researchers used an AlphaFold model to predict complexes made up of multiple proteins. In November 2021, an international group combined the strengths of AlphaFold and RoseTTAFold to evaluate interactions among 8.3 million pairs of proteins and predict large protein assemblies, involved in important biological functions, in the yeast Saccharomyces cerevisiae. Advances like these have also pushed the field further into protein design.
This is transforming much of medicine.
The ability to predict a protein’s shape computationally from its amino acid sequence – rather than determining it experimentally through years of painstaking, laborious and often costly techniques – is already helping scientists to achieve in months what previously took years.
The journal Nature has tracked the number of research articles, journal and preprint (those not yet peer reviewed). In December 2021 there were less than 10 research articles, But by March of 2022 there were well over 100.
“AlphaFold changes the game,” says Beck [molecular biologist from Max Planck Institute of Biophysics] . “This is like an earthquake. You can see it everywhere,” says Ora Schueler-Furman, a computational structural biologist at the Hebrew University of Jerusalem in Israel, who is using AlphaFold to model protein interactions. “There is before July and after.”
Another quote:
In the past half-year, AlphaFold mania has gripped the life sciences. “Every meeting I’m in, people are saying ‘why not use AlphaFold?’,” says Christine Orengo, a computational biologist at University College London.
Obviously, AlphaFold doesn't solve all problems. But it removes much of the laborious work that needs to be done to start researching a biological process. Intuitive leaps and medical trials must follow, but preliminary research has been hugely shortened. So too have clinical trials after COVID. Expect promising clinical trials later this year from protein research.
Or as Ark Invest analyst Alexandra Urman calculates things:
"Historically, the duration of pre-clinical testing from discovery to phase 1 has averaged four years," Urman wrote. "Based on our model assumptions, that number drops to three years and, with neural-network-based algorithms, two years...For the entire pipeline -- from discovery to the registration of a therapeutic drug -- technology could improve efficiency, reduce costs, and eliminate 3.5 years of research and development time."
Reduced costs and time have lured new companies and Universities into the field so we have more companies researching more issues with shorter timelines. Results should follow. Invest small amounts into promising candidates as research doesn't always work out and other companies may be working on better technology. But a few hits should more than make up for duds.
Please forward to interested third-parties. If you wish to have the regular Wednesday Musings in your email box, but aren’t ready for a paid subscription, just click on “Subscribe” and select “none”. If you are one of the few who wish to sponsor my effort, bless you. But you get the same messages as everyone.
I try to answer all questions or comments.