A friendly application for meta-analysis of genetic data in Info-Gen

BRUNO, C.; RUEDA CALDERÓN, A.; BALZARINI, M.

Barcelona

Conferencia; XXIXth International Biometric Conference; 2018

Info-Gen is a statistical software that provides a rich menu-driven interface and implements an accessible management that integrates a considerable variety of input (genotypic and phenotypic data) file formats. Furthermore, it allows to perform various analyzes of genetic diversity, to explore genomic data, and to study potential associations between molecular markers and a trait of interest. Its connection with R software has allowed an application to be added to perform Meta-Analysis (MA). MA is carried out on databases that collect results of genetic studies, either those produced with Info-Gen or those collected through Systematic Reviews. The MA application allows fitting statistical models of fixed and random effects to the compiled dataset to estimate the statistical significance of global genetic effects. Through MA, it is possible to improve the power and accuracy of the genetic effect (weighted average of the genetic effects reported in the primary studies). The data used to measure the effect of interest in each study, which will participate in the MA, can be continuous or discrete. Effect size estimation tools include statistics based on mean differences, standardized mean differences, ratio of means, odds ratios, relative risks, risk differences, proportions, and correlations. Additionally, the application allows to estimate statistics of heterogeneity between studies, to perform analysis by subgroups, and to carry out linear regressions and machine learning prediction models of the genetic effect on meta-covariables. It is particularly interesting when primary results are heterogeneous between studies. The application produces Forest-plot to visualize the results from primary association studies and the confidence interval of the overall effect through them. In this work, we illustrate an implementation of the MA application in Info-Gen to explore the consensus on the results of QTL (Quantitative Trait Loci) studies for disease tolerance in maize. An approachable protocol to perform MA in genome wide association studies is implemented.