Proteomics data sources.

HYE124: external benchmark

Two mixtures of human, yeast and E. coli proteins, measured in technical triplicates. Expected log₂(A/B) ratios are 0, +1 and −2 respectively. Download from the source and retain its sample metadata and processing provenance.

• Original PRIDE dataset: PXD002952
• LFQbench example files
• Study and methods

The Spectronaut example contains precursor measurements and needs documented protein-level summarization before matrix exploration. It is already normalized. Do not treat precursor rows as protein rows or absent measurements as zeros.

J-Express does not redistribute these files. The LFQbench repository states CC BY-NC-SA 4.0; the original PRIDE record states EBI terms of use. Follow the terms attached to the source you use.

CC0 candidate for a future demo

Proteomics in LKB1-null and -WT cells in 2D culture, John Ferrarone (2024), Harvard Dataverse V1. The dataset is marked CC0 and lists a MaxQuant proteinGroups.txt plus sample and TMT-label spreadsheets.

This is a TMT biological dataset, rather than a known-ratio LFQ benchmark. Download the protein report and both mapping spreadsheets together. Reporter channels, sample design, filtering and normalization need inspection before analysis. The protein report and both spreadsheets have been inspected: 10,030 protein groups, with 9,139 remaining after removing flagged decoys, contaminants and site-only groups. The mapping identifies ten samples (EV_1–EV_5 and WT_1–WT_5); six additional reporter columns are unmapped and must be omitted. Original intensities and zeros are preserved. This is a prepared reference candidate; an embedded demo and a validated inferential workflow have not yet shipped.