Australian researchers prioritized 32 genes with evidence suggesting causal roles in long COVID—13 of them novel—and identified three symptom-based biological subtypes. Using a novel MR-plus-Control Theory workflow combined with multi-omics data, the team made their platform publicly available for validation. Experts praised the approach but emphasized the need for laboratory validation, animal models and clinical trials before these findings can translate into therapies.
32 Genes Implicated in Long COVID and 3 Biological Subtypes Identified by Australian Researchers
The mystery of what causes the "brain fog," fatigue, chest pain and other symptoms of long COVID-19 has been traced to a set of 32 genes in an Australian study released this week. File Photo by geralt/Pixabay
Australian researchers report a major advance in understanding long COVID after prioritizing 32 genes they say have evidence consistent with causal roles in the condition, including 13 genes not previously linked. The findings, published this week in PLOS Computational Biology and Critical Reviews in Clinical Laboratory Sciences, also propose three symptom-based biological subtypes of long COVID that could help explain the condition's wide range of clinical presentations.
What Is Long COVID?
Long COVID is a multi-system disorder that can follow infection with SARS-CoV-2. Common manifestations include persistent fatigue, breathing difficulties, muscle and chest pain, dysautonomia (autonomic dysfunction affecting blood pressure and heart rate), and cognitive problems often described as "brain fog." Symptoms can last for months or years after the acute infection. Estimates of prevalence vary widely (roughly 10%–70% in different studies), with a global burden estimated at about 65 million people.
How The Study Was Done
The research team, largely from the University of South Australia, combined two advanced genetic approaches—Mendelian randomization (MR) and network Control Theory (CT)—and integrated them with large-scale multi-omics data (genomic, epigenomic, transcriptomic and proteomic datasets). This MR-plus-CT workflow with adjustable weighting was used to prioritize gene candidates and strengthen evidence for causality rather than mere association.
"We prioritized 32 candidate genes with evidence consistent with causal roles, including 13 not previously linked — which is more actionable for therapy development than correlation alone — while still requiring experimental follow-up," said Sindy Licette Piñero, a doctoral candidate and co-author.
Key Findings
The study:
- Prioritized 32 genes with evidence suggesting causal involvement in long COVID, including 13 novel candidates.
- Identified three symptom-based biological subtypes of long COVID, each with distinct clinical and molecular signatures.
- Released the analytical platform publicly to enable validation and extension across other cohorts.
Expert Reactions and Next Steps
Independent researchers praised the methodology and the prioritized gene list but urged caution about immediate clinical implications. They stressed that experimental validation—animal models, laboratory studies and eventually clinical trials—will be required to determine whether targeting these genes can prevent or treat long COVID.
"While the approach is statistically advanced and helps prioritize causal genes, interpretability is hampered by methodological complexity," said Art Schuermans of KU Leuven and the Broad Institute.
The authors acknowledge the need for follow-up experimental work to confirm biological mechanisms and to assess the clinical actionability of the candidates. If validated, the gene list and subtype framework could guide the development of targeted therapies and personalized management strategies for people living with long COVID.
Bottom Line
The study marks an important step toward disentangling the heterogeneous biology of long COVID by proposing plausible causal genes and biological subtypes, while emphasizing that further validation is essential before these findings can inform treatments or diagnostics.
