Maeve McGillycuddy

Phylogenetic generalised linear mixed models (PGLMMs) help ecologists to distinguish ecological drivers from other processes shaping evolutionary patterns, yet existing implementations are often limited in distributional scope or computational speed. We compare five R packages for fitting PGLMMs and highlight the new covariance structure propto in the general-purpose GLMM package glmmTMB . Simulations show that glmmTMB fits PGLMMs faster overall than brms , MCMCglmm , INLA , and phyr , while producing similar model estimates. We present the first practical application of glmmTMB for fitting phylogenetic random effects using likelihood-based models that accommodate repeated measures, demonstrated through case studies of evolutionary trait data. By improving both speed and flexibility, glmmTMB broadens access to PGLMM and supports deeper insights into trait evolution and diversification.

PURPOSE: Patient-derived xenograft (PDX) models are used in precision medicine to identify treatments that may be more effective for an individual patient's molecular tumor profile. We explored parents' perspectives of using personalized PDX mouse models to potentially guide treatment recommendations for their child enrolled in Precision Medicine for Children with Cancer (PRISM), a precision medicine trial for poor prognosis cancer. METHODS: We report on interview data collected from a mixed-methods study examining families' experiences of PRISM. We interviewed 57 parents about their understanding and perceptions of the use of PDXs in PRISM. We analyzed quantitative data using linear models and Fisher-Freeman-Halton exact tests and qualitative data via inductive content analysis. RESULTS: Overall, 83% of participants deemed PDXs acceptable. This is despite 51% of participants being aware that PDXs were a part of PRISM and the majority having little-to-no understanding of the technology. Of the 14 potential advantages/disadvantages of PDXs presented, participants rated that “accessing the right drug more quickly” as most important and “testing will involve harming animals” as least important. Several parents had concerns around animal models although some preferred mouse models over zebrafish as they felt that mice were more human-like and thus more valid. Participants expressed that they would sacrifice an unlimited number of mice for their child, yet only 63% endorsed a willingness to use up to 1,000 mice per patient. Willingness to use more mice appeared to be associated with parents' sense of urgency for a cure and perceived likelihood of benefit for their child. Regardless, most participants (96%) indicated that the number of mice used in PRISM would not have affected their decision to participate. CONCLUSION: Our study indicates that despite having minimal knowledge, parents consider the use of PDX mouse models in precision oncology to be acceptable.