Software

MARTin: Automating Data Preparation and Parameter Specification for BLUPF90 Analyses

---

Abstract

The objective of this study was to develop an automated pipeline for preparing datasets and parameter files for BLUPF90 in multi-trait genetic evaluations, where manual processing is time-consuming and prone to specification errors. Preparing BLUPF90 inputs for multiple traits, covariates, and model components typically requires extensive manual coding, which can introduce inconsistencies and reduce the reproducibility of genetic evaluations. To address these limitations, we developed MARTin, an R-based pipeline designed to automate data preparation, effect specification, and the construction of BLUPF90-compliant parameter files along with their associated phenotype files. MARTin consists of three sequential modules. Module 1 extracts descriptive statistics and distribution summaries generated through the summarytools R package. Module 2 retrieves model specifications derived from regression-based variable selection performed with the OLSRR package. Module 3 uses this information to construct the full set of BLUPF90-compliant parameter files, generating independent folders for each trait to enable efficient Linux-based batch execution. The pipeline is compatible with single and bi-trait analyses under animal, maternal and permanent environmental models and uses CSV-formatted data as input. Performance was evaluated using simulated datasets comprising 75,000 phenotypic records for 24 traits and a pedigree of 150,000 animals. MARTin successfully produced all BLUPF90-compliant files, accurately mapping effects, factor positions, and model components without manual intervention. Parameter-file construction was completed within seconds on a standard workstation, and repeated executions (R = 5) yielded identical outputs, confirming full reproducibility. Computational performance remained stable when scaling from single- to multi-trait analysis. In conclusion, MARTin provides a reliable, scalable, and reproducible solution for preparing datasets and parameter files for BLUPF90-based genetic evaluations. By automating error-prone pre-analysis steps, it reduces manual inconsistencies, improves traceability, and supports more efficient multi-trait genetic-evaluation pipelines.The MARTin pipeline will be released as an open-source software repository on GitHub, with documentation and example workflows provided to support reproducibility.

Keywords: 2026

How to Cite:

Ramirez-Diaz, J., Rocha-Silva, M., Cozzi, P., Stella, A. & Biffani, S., (2026) “MARTin: Automating Data Preparation and Parameter Specification for BLUPF90 Analyses”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286834. doi: https://doi.org/10.31274/wcgalp.24139

Rights: 1