We studied mouse cardiac fibroblasts in the first days after myocardial infarction (MI). Col1a1-GFP reporter mice (C57BL/6 background, 8-10 weeks old) received an MI by permanent ligation of the left anterior descending coronary artery, and their hearts were harvested 1, 2, 3, 4, 5 or 6 days later; from each harvest we FACS-sorted Col1a1-GFP+ CD31- CD45- cardiac fibroblasts. The six harvest days are the groups we compare. They are recorded in the `group` column, from 'Col1a1 GFP+ Cardiac Fibroblasts 1 days after MI' to 'Col1a1 GFP+ Cardiac Fibroblasts 6 days after MI', with 3, 3, 4, 3, 1 and 3 samples for days 1 to 6 (17 in all). The day appears only inside these labels; there is no separate time column. All mice had an MI, and there is no healthy or sham group.

Mice are killed at harvest, so each sample is a separately sorted cell population and no mouse appears at two harvest days; whether a sample came from one heart or from several pooled hearts is not recorded. Each library was sequenced once, so there are no technical replicates. The MARS-seq library pools and sequencing runs were not recorded, so we have no batch information. Mouse IDs and sex are not recorded either, and no other variable matters to us; the remaining columns are constant annotations or sample names.

The data are bulk RNA-seq made with the MARS-seq protocol. The matrix holds raw gene-level read counts from featureCounts, one row per Ensembl gene ID (mm10).

Our question is which genes differ in expression among the six harvest days: an overall difference anywhere among the six `group` levels, with day 1 as the baseline, rather than a comparison of two particular days. All 17 samples, including the single day-5 sample, are part of this question.