This person is a member of Sanger Institute Alumni.
I am a biostatistician supporting the International Mouse Phenotyping Consortium by researching methods to optimise the experimental design and data analysis. My focus has been on developing robust statistical analysis pipelines that consider the complex data and variation in workflow within phenotying experiments.
My role requires cross discipline collaborations to address the complex data which has challenges that arise in identifying reproducible phenotypes and those that arise from the high throughput nature of the project. The research looks to understand the issues and identify solutions that will improve the ability of the experiments to answer the research question.
There have been published concerns over replicability of phenotyping studies. To address issues with replicability I have focused on two angles.
- Firstly to optimise the data analysis methods applied. The high throughput pipelines provide a unique opportunity to explore large datasets and then develop and test various statistical methods using resampling and simulation studies. The methods I have developed have been made freely available to all through the development of an R package called PhenStat.
- The second approach is looking at how we can improve the design and reporting of the experiment conducted. To improve reporting, I led the development of the Mouse Experimental Design Ontology to provide a standard language to capture how we conduct the experiments. Across the consortium, we have lead the field, and applied the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines to a database. Following my involvement in the development of the NC3Rs Experimental Design Assistant, we are testing programs to introduce the software at the institute.