Datasets and R Markdown files for the article “LabRI: a data-source-agnostic, multi-method consensus framework for laboratory reference interval estimation” submitted to Clinica Chimica Acta
Description
This repository contains the datasets and R Markdown scripts supporting the article "LabRI: a data source-agnostic, multi-method consensus framework for laboratory reference interval estimation" submitted to Clinica Chimica Acta. The repository includes the Group 1 (real-world) and Group 2 (simulated) datasets; the ZEBRA1.Rmd and ZEBRA2.Rmd scripts implementing the Z-score Evaluation, Benchmarking, and Ranking Algorithm for method comparison through standardized z/zlog scores, Lin's concordance correlation coefficient, ordinal scoring, and ranking; the DataSampleR script for reproducible random subsampling; Word reports with reference-interval estimates obtained with TMC, TML, reflimR, and refineR via the ReferenceRangeR web tool for all evaluated datasets; and the complete outputs generated by ZEBRA and DataSampleR.This repository contains the datasets and R Markdown scripts supporting the article "LabRI: a data source-agnostic, multi-method consensus framework for laboratory reference interval estimation and verification" submitted to Clinica Chimica Acta. The repository includes the Group 1 (real-world) and Group 2 (simulated) datasets; the ZEBRA1.Rmd and ZEBRA2.Rmd scripts implementing the Z-score Evaluation, Benchmarking, and Ranking Algorithm for method comparison through standardized z/zlog scores, Lin's concordance correlation coefficient, ordinal scoring, and ranking; the DataSampleR script for reproducible random subsampling; Word reports with reference-interval estimates obtained with TMC, TML, reflimR, and refineR via the ReferenceRangeR web tool for all evaluated datasets; and the complete outputs generated by ZEBRA and DataSampleR.
Files
Steps to reproduce
𝟭. 𝗦𝘆𝘀𝘁𝗲𝗺 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝗺𝗲𝗻𝘁𝘀 Install the R environment and IDE before using the materials in this dataset: - R (version 4.6.0 or compatible) - RStudio (version 2026.04.0+526 or compatible) Note: A desktop workstation with at least 16 GB of RAM is recommended. 𝟮. 𝗗𝗼𝘄𝗻𝗹𝗼𝗮𝗱 𝗮𝗻𝗱 𝗲𝘅𝘁𝗿𝗮𝗰𝘁 - Download the dataset from Mendeley Data and extract the .zip archive to a local working directory. Preserve the folder structure; do not move files between folders. 𝟯. 𝗜𝗻𝘀𝘁𝗮𝗹𝗹 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝗱 𝗥 𝗽𝗮𝗰𝗸𝗮𝗴𝗲𝘀 Each tool folder in the dataset contains a dedicated Install_packages.Rmd script that installs all R packages required by that tool. Open the Install_packages.Rmd of each folder in RStudio and run it once before executing any analysis script in that folder. 𝟰. 𝗥𝗲𝗽𝗿𝗼𝗱𝘂𝗰𝗲 𝘁𝗵𝗲 𝗮𝗻𝗮𝗹𝘆𝘀𝗲𝘀 Before executing each tool, verify that all required input files are placed in the expected folders and that the user-defined fields within the scripts are properly filled with the specific data for the analysis. The materials execute end-to-end and generate all tables, figures, and outputs reported in the article: - ZEBRA1.Rmd and ZEBRA2.Rmd: reference-interval method comparison via the ZEBRA framework. Open the R Markdown file in RStudio and run all chunks sequentially, or knit the document. - DataSampleR.Rmd: reproducible random subsampling of datasets. Open the R Markdown file in RStudio and run all chunks sequentially, or knit the document. - LabRI Shiny application: interactive interface for reference-interval estimation and verification described in the article.