A Practical Precision Medicine Approach to Complex Chronic Diseases
By Deborah Borfitz
June 3, 2025 | Although not often counted among the major chronic diseases, myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) and long COVID are both highly prevalent and uniquely challenging because the underlying causes are not fully understood, treatment needs to be personalized, and the often-invisible nature of the symptoms can be stigmatizing. Long COVID alone—barely a recognized disease five years ago—affects about 400 million patients’ lives and has an annual economic impact equivalent to about one percent of the global economy (Nature Medicine, DOI: 10.1038/s41591-024-03173-6).
An astounding 250 symptoms affecting most of the major organs of the body have been linked to long COVID. That unsettling list has some crossover with the experience of ME/CFS patients given that the two conditions are genetically related to one another, says Steve Gardner, Ph.D., founder and CEO of UK-based PrecisionLife.
Scientists at PrecisionLife were responsible for identifying the first 14 genes associated with ME/CFS using the predominantly white European ancestry UK Biobank populations as well as the 73 associated with long COVID in the Sano GOLD dataset (from Sano Genetics) with a similar representation bias. They subsequently looked for genetic overlap and found nine genes common to the two disease populations.
Digging in further, the PrecisionLife team most recently demonstrated that seven of those nine genes could be reproduced in a long COVID population pulled from the independent and ancestrally more diverse All of Us dataset of the National Institutes of Health (Journal of Translational Medicine, DOI: 10.1186/s12967-025-06535-x). The study also confirmed disease signatures associated with 11 of the 13 drug repurposing candidates identified in the original Sano GOLD study, supporting the development of both biomarker-driven diagnostics and targeted treatments.
The enabler of all this was PrecisionLife’s combinatorial analytics platform that looks beyond a single mutation inside of a gene that leads to a cascade of problems. That’s a great approach for classic diseases like Huntington’s, cystic fibrosis, or sickle cell. But complex chronic diseases like ME/CFS and long COVID are not like that, Gardner says, so single genetic associations can’t adequately explain them.
That the seven causal genes could be reproduced in the All of Us dataset in the latest study indicates researchers are “on the brink of understanding the complex biology that underpins patients’ likelihood to go on and develop [long COVID or ME/CFS] and the mechanisms behind that susceptibility,” says Gardner. Doing that analysis in diverse populations ensures that the identified mechanisms apply to the most patients possible, he adds, noting 55% to 75% of the disease signals replicate in Black and Hispanic populations.
Based on unpublished data, the “strong expectation” of investigators is that all nine of the genes thought to be common to ME/CFS and long COVID will be a reproducible finding. The All of Us dataset used in the latest study had under 500 patients while an additional 1,500 patients populated the most recently released cohort of data, Gardner says, which will also permit a more detailed examination of reproducibility in general.
This soon-to-publish confirmatory study will be particularly significant for the ME/CFS community, which has long struggled to have their disease taken seriously by many health practitioners. “This is putting [ME/CFS] on a footing with and actually connecting it to the research going on with long COVID,” he says.
Precision medicine has traditionally been pointed at oncology and rare diseases, but many diseases outside those therapeutic arenas are more multifactorial, Gardner stresses. “We are starting to understand how those diseases work and get deep insights... in a way we couldn’t do 10 years ago.” The raft of new opportunities this presents comes from the confluence of a lot of great data, a large amount of compute capacity, and advanced analytical techniques like smart maps powered by artificial intelligence (AI).
Minding the Gap
The long-term objective here is to tackle the 40% to 45% of disease signals that can’t be explained by single genetic associations, says Gardner. A complex chronic disease may have a heritable component of 50%, but a single gene may account for only 5% to 10% of the disease. The gap, he explains, lies in the “non-linear interactions between multiple genes” that happen via metabolism and amplifies or inhibits each other’s effects.
“You can’t reconstruct that by looking at one gene at a time,” Gardner adds. “You have to look at the whole system.” PrecisionLife isn’t alone in making that intellectual leap, but it is the only precision medicine company to come up with a practical algorithm that can do the analysis across populations.
Its combinatorial analytics platform underpins everything PrecisionLife does, and to date the AI-powered technology has been applied to about 60 different chronic diseases, he reports. The novel AI tools being employed are the brainchild of Gert Møller, Ph.D., one of the company’s cofounders who serves as chief analytics officer.
The platform identifies mechanistic patient stratification biomarkers that can be translated into “Mechanostics” tests for clinical use to enable more personalized treatments, says Gardner. PrecisionLife’s Mechanostic test for ME/CFS captures 199 unique single nucleotide polymorphisms (SNPs) and can predict disease “in the ballpark” (odds ratio of 4.5) with BRCA mutations (odds ratio of 5 to 7.5) for breast cancer, which is “unexpected in a disease with no prior genetic associations.”
The platform is easy to use because genotyping is an established and widely available practice and the Mechanostics tests under development are designed to be conveniently and economically administered at home with the buccal (cheek) swab samples sent to a lab for processing, Gardner says. “If you know what’s driving the disease you can choose what you want to look for” to gain insights about those etiologies, the risk of one disease versus another, and the drugs most likely to work for individual patients.
Novel Drug Targets
The combinatorial analytics approach was first developed in 2017, and one early target was amyotrophic lateral sclerosis (ALS), also known as Lou Gehrig's disease, says Gardner. Notably, PrecisionLife identified 24 key genes linked to “actively protective biology,” including the gene responsible for removing the neurotoxin homocysteine. This partially underpins last fall’s market approval in Japan of mecobalamin, a high-dose form of B-12 found to significantly slow disease progression, for pharmaceutical company Eisai.
This is but one example of “a completely new set of drug targets” that medicine developers could be going after, potentially with benefits for everybody who has a particular disease if the resilience mechanisms are augmented and supported, he continues. “It doesn’t matter how you came to have the disease; that resilience mechanism is going to work as well to improve your experience, reduce the symptoms, and hopefully reduce progression of the disease.”
PrecisionLife has done a lot of work identifying novel targets and biomarkers for pharmaceutical companies but has been pivoting to the development of clinically actionable diagnostics, says Gardner, since few great options currently exist for many complex chronic diseases. So, what the company does now is turn the insights it generates into buccal swab tests that involve collecting DNA for genotyping.
Critically, these Mechanostics tests get at the mechanisms driving the disease inside an individual since “many of these are very heterogenous diseases and patients will not respond to all the drug options that may be available,” Gardner says. As with companies like Tempus AI, PrecisionLife leverages large datasets and AI, but its combinatorial approach to finding disease signals is focused on complex chronic diseases rather than oncology.
Current research efforts around long COVID are many and varied, including studies based on the hypothesis that viral persistence plays a role, as it no doubt does, says Gardner, at least for some people. The effects on mitochondria have also been witnessed, as well as immune issues that possibly determine how a disease manifests and transitions from a primary viral infection into a more persistent stage with long-term inflammation and significant impacts on the vasculature.
“But we also see really strong metabolic signals inside the disease,” he continues, including the insulin receptor regulating blood sugar levels and overall metabolism, and the clock gene, responsible for circadian rhythm. Complicating matters is that each of these “symptomatic buckets” can individually affect 30% to 40% of patients.
Parallel Studies
PrecisionLife launched two parallel studies in the UK and the U.S. that are shedding light on ME/CFS and long COVID and pointing to new diagnostic and therapeutic options, says Gardner. First up in December 2023 was the LOCOME project, funded by Innovate UK, which is examining a dataset of 25,000 ME/CFS patients from the DecodeME study developed by Action for ME and the human genetics unit at the University of Edinburgh, together with long COVID cohort from All of Us. Genomics, phenotypic, and relevant demographic data are part of the ongoing analysis.
The initial reproducibility analysis happened more quickly than planned, as did the search for a clinical validation partner, he adds. Researchers landed on the Metrodora Foundation (now called the Complex Disorders Alliance), whose associated institute was running a clinic for ME/CFS and long COVID patients and provided access to those populations for recruitment into the so-called MELO (ME and Long COVID Observational) trial that launched last year.
A series of clinical studies are planned for 1,000 patients, half with ME and half with long COVID, says Gardner. One aim is to evaluate the diagnostic accuracy of the Mechanostic tests and their ability to identify people at high genetic risk of disease development. Participants are being given results indicating their specific disease risk signatures and therefore the symptoms they are most likely to experience.
Another more ambitious goal is to search for repurposing treatments for ME and long COVID as quickly and cost effectively as possible, Gardner says. Participants are patients from the main Mechanostics trial with specific mechanisms driving their disease and a safe, well-tolerated generic treatment already identified.
The LOCOME and MELO studies “are feeding results and information to each other and accelerating the process,” he says. That means the Mechanostic test and repurposing studies should all be completed within two years—almost twice as fast as originally predicted.
Women’s Health Issues
Given that females significantly outnumber males in terms of ME/CFS and long COVID prevalence, both should be considered important women’s health issues, Gardner says. The diseases are remarkably similar and many of the afflicted patients tend to be relatively young women.
“We also see a lot of comorbidities that also relate to diseases that are known to have a strong female bias,” he continues, specifically mentioning migraines, Ehlers-Danlos syndrome, and hypermobility syndrome. One of the studies PrecisionLife has underway is looking at those comorbidities and the overlap between them.
The company has extensively studied endometriosis, adenomyosis and related women’s health disorders where, again, several of the same genes show up—particularly those associated with the debilitating pain patients tend to experience. In the most severe forms of these conditions, many women also become oversensitive to light, noise, and movement possibly because of other types of mechanistic factors, says Gardner. “Understanding that gives us the opportunity to go after these core early drivers and symptoms of the disease, and if you can offer a drug for one disease the expectation is it may well be useful across multiple indications.”
Gardner says he is especially optimistic about the role of protective signatures relative to what has genetically gone wrong, which has been the basis of nearly the entire drug industry up to now. “We know that biology loves homeostasis, it loves preserving the normal functioning of cells in bodies, so surely there are mechanisms that are resisting disease pressure,” he reasons.
To that end, PrecisionLife has been actively looking for those resilience mechanisms in healthy control populations that genetically have all the same risk factors and triggers as patient cohorts and yet never get the expected disease. “We see [this phenomenon] in every disease that we study.”
PrecisionLife just released a report on a study finding many protective signatures mapping to nine protein-coding genes associated with ME/CFS and several more in ALS (Artificial Intelligence in the Life Sciences, DOI: 10.1016/j.ailsci.2025.100125). The research team is now doing another study in long COVID, which shares some of those same genes, and looks to publish those results soon.