A repository to hold the data, code and results of the paper "A consensus score to combine inferences from multiple centres"
BiorXiv: https://www.biorxiv.org/content/10.1101/2022.11.10.515944v2
- Hamed Haselimashhadi hamedhm@ebi.ac.uk
- Kolawole Babalola
- Robert Wilson
- Tudor Groza
- Violeta Muñoz-Fuentes
Experiments in which data are collected by multiple independent resources, including multicentre data, different laboratories within the same centre or with different operators, are challenging in design, data collection and interpretation. Indeed, inconsistent results across the resources are possible. In this paper, we propose a statistical solution for the problem of multi-resource consensus inferences when statistical results from different resources show variation in magnitude, directionality, and significance. Our proposed method allows combining the corrected p-values, effect sizes and the total number of centres into a global consensus score. We apply this method to obtain a consensus score for data collected by the International Mouse Phenotyping Consortium (IMPC) across 11 centres. We show the application of this method to detect sexual dimorphism in haematological data and discuss the suitability of the methodology.