Lipomas are clonal neoplasms, meaning they arise from a single cell that began dividing uncontrollably but never acquired malignant properties. Alongside this work, the archive holds studies that examine tumors from various angles: how genes predict breast cancer outcomes, how a tumor builds its own family tree, how blood reveals a tumor's fate. Our work sits off to the side: it is not about a malignant process but about a benign tumor, and not about somatic mutations but about inherited predisposition. This is a different scale: not the fate of a tumor that already exists, but the probability of one appearing. Lipomas were once thought to arise from random rearrangements in cells, but family and syndrome data hinted at a hereditary component. Now population genomic resources allow this to be tested directly.
The author took FinnGen data on 13,526 lipoma cases and 486,660 controls, ran a genome-wide association search and fine-mapping with SuSiE and FINEMAP to isolate independent signals. He then annotated all variants within the 95% credible sets without looking at the telomere results, and found that genes in these loci often participate in telomere maintenance and DNA repair. Next came Mendelian randomization: he took 210 variants associated with leukocyte telomere length from All of Us and tested whether they were linked to lipoma risk in FinnGen. The analysis was repeated with independent telomere length effects from UK Biobank, 15 variants in canonical telomere genes were excluded, individual anatomical sites of lipomas were checked, and the local genetic architecture was examined in three key regions: TERC, TERT, and STN1. An agentic workflow involving ChatGPT and Codex was used, but all scientific decisions were made by a human.
The genome-wide analysis revealed nine independent signals in seven regions. The most precise signal on chromosome 5 consisted of a single variant with a PIP of 0.997, whereas the signal on chromosome 22 contained 145 variants and was poorly resolved. Annotating the loci without telomere data pointed to the genes TERT, RTEL1, STN1, and TERC, as well as to processes of telomere elongation, maintenance, and replicative aging. Mendelian randomization showed a strong association: genetically predicted longer telomeres increased lipoma risk (β=0.445, 95% CI 0.355–0.535, P=2.49×10⁻²²). The association replicated with UK Biobank (β=0.587, 95% CI 0.459–0.715, P=3.09×10⁻¹⁹), persisted after excluding canonical telomere genes (β=0.439, P=1.06×10⁻¹⁶), and was observed for lipomas of the limbs, trunk, and head/neck. Locus-level analysis revealed heterogeneity: TERC showed strong support for a shared signal (PP.H4=0.989), TERT showed strong overlap but different candidate variants, and STN1 showed a clear distinction between signals (PP.H3=0.9995).
What sets this work apart from its neighbors in the archive is that it treats a benign tumor as a model for studying the trade-off between regenerative capacity and protection against neoplasia. Neighboring studies focus on malignant tumors and their evolution, whereas here it is shown for the first time that inherited telomere length influences the probability of a benign clonal expansion specifically. This broadens the picture of telomeres' role: they do not only limit cancer, they also modulate the risk of harmless yet clinically noticeable neoplasms. The result also underscores that a shared genetic signal between two traits does not mean the same mechanism at every locus.
The field could move toward integrating inherited genetics with the somatic genomes of individual lipomas. If germline variant data and tumor mutations are combined, one can test whether a predisposition to long telomeres affects the likelihood that specific somatic events take hold. It would also be promising to study whether the effect differs in people with single versus multiple lipomas, and to extend the analysis to other benign mesenchymal tumors, such as uterine leiomyomas.
The findings could influence genetic counseling for multiple lipomas and the understanding of side effects from therapies aimed at telomere elongation. They are also useful for interpreting genome-wide associations with other benign neoplasms.
The immediate next step is sequencing paired germline and tumor samples from people with single and multiple lipomas to test how telomere background influences clonal evolution. Larger lipoma GWAS and data on regulatory elements in adipocytes are also needed to pinpoint the causal genes.
In the genetics of benign tumors, it remains unresolved why some people are prone to multiple lipomas and others are not, and which hereditary factors govern the transition from normal fatty tissue to clonal growth. This work directly links that problem to telomere biology, but it does not explain why the effect varies across anatomical sites or why some loci, such as STN1, show no shared signal.
In a series of 188 tumors, only one of 58 lipomas from patients with multiple lipomas had detectable chromosomal changes, which led scientists to suspect a hereditary predisposition as early as 1988.
In Neal Stephenson's novel The Diamond Age, genetic manipulations of telomeres are used to extend life but lead to unforeseen consequences. This work shows that telomere elongation may come at a price in the form of increased tumor risk, albeit benign ones.
Not published on arXiv. Here are its page, the PDF and the formalization: you can check it yourself, not only read our account.