Associations between neoplasia incidence and telomere biology
We compiled neoplasia incidence rates in 7 species of mammals from a medically curated postmortem database published by Boddy et al. (2020) in addition to 15 species from various sources in the literature reviewed by Albuquerque et al. (2018) and shown in Table S10. In all studies, the number of individuals with abnormal growth including both benign and malignant tumors (neoplasia) out of the total number of necropsies were used to define the neoplasia prevalence rate (%). For the large treeshrew (Tupaia tana ) neoplasia rate is unknown, so we used the estimate reported for the closely related common treeshrew (Tupaia glis ). Due to small samples sizes, neoplasia data were combined in Boddy et al. (2020) for the Asian and African elephant (Elephas maximus and Loxodonta africana ), so we used the average TL, mass and lifespan of the two species, which are similar (Table S9), in subsequent analyses. Boddy et al. (2020) also provided estimates of lifetime prevalence of malignant neoplasms (invading surrounding normal tissue) of 29 mammal species, allowing us to test if neoplasia prevalence is generally a good predictor of malignancy prevalence (i.e. cancer rate in %). Most neoplasia data are from wild animals kept in captivity with no age or sex information, and not all tissues were investigated during necropsies, thus the data only provide crude estimates of the natural rate of spontaneous cancers.
Species mean telomere length measured using the TRF method allows comparisons of absolute TLs and were collected from both Gomes et al. (2011) and Seluanov et al. (2007) to increase sample size (Table S10). Six overlapping species in these two studies show that the TL measurements are highly correlated (R²=0.9431), but consistently higher in Seluanov et al. (2007). We therefore used an ordinary linear regression of TLs of the overlapping species to adjust measurements of the four species (Table S10) from Seluanov et al. (2007) to match Gomes et al. (2011) using the formula: 0.6617 x TL - 2.4454 kb. Telomerase activity was measured using different methodologies, so we can only report whether telomerase was present (1) or absent (0) in the somatic cells of the species. We then test phylogenetic bivariate associations between neoplasia rate and TL, telomerase expression, adult body mass and maximum lifespan (obtained from the AnAge database, Tacutu et al., 2018), and whether TL and/or telomerase expression predict neoplasia rates (response variable) using phylogenetic multiple regression as described above.