About my research
I'm a statistician* working in the Institute of Biodiversity, Animal Health and Comparative Medicine (IBAHCM), in the University of Glasgow. My main interests are in analysing genetic data, and GLMMs. I also spend a lot of time helping other scientists use data to find reliable answers to their research questions.
Applied research:
SEEDZ (2016-): Social, Economic and Environmental Drivers of Zoonoses Livestock, Livelihoods and Health. My role on the project is to model zoonosis transmission in Northern Tanzania (SEEDZ modelling PIs: Rowland Kao & Louise Matthews)
Epi-SEQ (2013-16): Molecular epidemiology of epizootic diseases using next generation sequencing technology (Glasgow PI: Dan Haydon).
AvecNet (African Vector Control: New Tools, 2013-16), a collaborative project between African and European Researchers to develop and evaluate new tools for malaria control in Africa (Glasgow PI: Heather Ferguson).
ECOTEL (2012-15): The ecological significance of telomere dynamics: environments, individuals and inheritance (PI: Pat Monaghan).
Investigating the role of HLA alleles in causing Hodgkin lymphoma (PI: Ruth Jarrett).
Methodological research:
Helping to design experiments for the AvecNet project got me interested in using simulations to do power analysis for generalised linear mixed models (GLMMs). The resulting paper and tutorial can be found here. I've made the simulation function, plus some R functions for assessing the fit of GLMMs, available in an R package, GLMMmisc.
I published a note showing how to account for random slopes when estimating R2 for GLMMs, based on Nakagawa & Schielzeth's R2GLMM method. This method has been incorporated into the piecewiseSEM and MuMIn R packages, and featured in this blog post. I recently contributed to further extending Nakagawa & Schielzeth's R2GLMM to a wider range of GLMMs.
In collaboration with Dan Haydon I wrote Pedant, a program for estimating microsatellite genotyping error rates.
PublicationsGoogle Scholar
SoftwarePedant, a program for estimating microsatellite genotyping error rates.
GLMMmisc, an R package with various useful functions for GLMMs.
TeachingA lecture in three parts on practical experimental design delivered to masters students on the the IBAHCM Key Research Skills course:
KRS Experimental Design 2.1: Reverse planning (12 minute video lecture)
KRS Experimental Design 2.2: Reverse planning examples (12 minute video lecture)
KRS Experimental Design 2.3: Power analysis (49 minute video lecture)
Other activities
I host Statgen, which is a mailing list devoted to statistical genetics.
Contact
Institute of Biodiversity, Animal Health and Comparative Medicine
Graham Kerr Building
University of Glasgow
Glasgow G12 8QQ
UK
University of Glasgow webpage: http://www.gla.ac.uk/bahcm/staff/pauljohnson
Email: pcdjohnson [at] gmail [dot] com
Twitter: @paulcdjo
*"A statistician is one who prefers true doubt to false certainty" - Stephen Senn