Summarize this content to 2000 words in 6 paragraphs in Arabic Unlock the Editor’s Digest for freeRoula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.Proteomics — the study of the proteins that are the building blocks of life — has emerged as a new digitally-powered weapon in scientists’ efforts to better understand and treat diseases. A project launched this year by the UK Biobank genetic database and 14 drugs companies aims to build on breakthroughs already achieved in the diagnosis of cancers, autoimmune conditions and dementia.The proteomics initiative is a flagship example of how advanced computers and artificial intelligence models can harness big biological data sets to look deeper into how the human body works and can malfunction. Proteomics’ potential lies in the extra level of sophistication it offers: it allows scientists to move beyond analysing genes to examining the proteins whose production the genes instruct.A central aim is to use the size of the UK Biobank data set to train AI models to identify disease subtypes more precisely, allowing treatments to be tailored and timed for maximum effectiveness. “[Researchers] will be able to look to see how lifestyle, environment and genetics lead through proteins to some people developing particular diseases and others not,” says Professor Sir Rory Collins, UK Biobank’s principal investigator and chief executive. “We can then look at ways in which to prevent those conditions before they develop. And also the proteins . . . will help us to identify new targets, new ways in which to treat disease.”If genes are the plans needed for construction of the human body, proteins are the materials used to build it. Proteomics enables scientists to look closer than ever before at the assembly process. It should enable them to see crucial details such as whether some genes are present, but have their instructing function switched off. The UK Biobank, which was set up almost 20 years ago, is the world’s leading genetic research database because of its combination of size and longevity. The older such an information trove is, the more analytical value it has — because more time has elapsed to observe how diseases appear and progress in its human cohort.The Biobank will deploy its genetic information from half a million people and combine this with secondary samples taken from 100,000 of those volunteers up to 15 years later. Together, these will enable scientists to monitor the impact on disease of changes in the levels of up to 5,400 proteins in mid- to late-life.The consortium of pharmaceutical companies funding the project includes some of the industry’s biggest names. It comprises Alden Scientific, Amgen, AstraZeneca, Bristol Myers Squibb, Calico Life Sciences, Roche, GSK, Isomorphic Labs, Johnson & Johnson, MSD, Novo Nordisk, Pfizer, Regeneron and Takeda. They will have access to the results nine months before they are published to researchers worldwide. The industry and academic researchers are excited by the findings of a pilot Biobank proteomics project that released results in 2023, which analysed almost 3,000 proteins in 54,000 participants. The study allowed researchers to make more than 10,000 previously unknown links between common genetic variants and changes in protein levels. Scientists have used the results to improve the prediction of disease and to target treatments for breast cancer, cardiovascular disease, Parkinson’s disease and other brain conditions.The new initiative is almost 10 times larger in scale. It offers the potential to gain insights by cross-referencing Biobank’s genetic data with information from other sources. One would be magnetic resonance images taken of the brains, hearts and other parts of the body of 100,000 of the Biobank participants. The project is expected to drive the development of improved artificial intelligence models to better deal with diseases, in part by investigating how protein levels change with disease and over time. These would build on existing uses of the technology that help predict the likelihood of some conditions by using biomarkers, imaging scans or genetic risk factors. The study of proteins has been greatly boosted by advances in analytical technology over the past few years. One is the AlphaFold AI model developed by Google DeepMind for protein structure prediction. The work, which led in 2022 to the most complete and accurate database yet of almost every known protein, won DeepMind’s Sir Demis Hassabis and John Jumper one half of last year’s chemistry Nobel Prize. AlphaFold launched its third iteration last year with the promise that it would allow more detailed exploration of how biochemical networks involving proteins work to sustain our bodies’ cells. Hassabis described the AlphaFold innovation last March as a more efficient way “to search for the needle in a haystack”.AlphaFold 3 extends its analysis to the DNA and RNA genetic codes as well as ligands — molecules that bind to others and can be important markers of disease.AlphaFold and other developments promise to turbocharge long-running efforts such as the Human Proteome Project that was launched in 2001 and brands itself as “translating the code of life”. This international research collaboration seeks to find and identify all the proteins in the human body. By 2023, it claimed to have found 18,397 of 19,778 predicted proteins encoded by the human genome, or 93 per cent of the total. The scale of the proteomics endeavour has telling parallels to the quest completed in 2003 to sequence the human genome. The computer-assisted study of genetics has driven many health discoveries since. The ambition now is to further that revolution by applying still more digital techniques to the biologically fundamental world of proteins.
