STDetail example files

expression-tables/
  measured-rna.csv                    Measured RNA of HEST section TENX118 (lung): one row per spot,
                                      sample_id, unit_id, x, y, then 50 genes as log1p(raw count).
  Standard training.csv               Ridge-Phikon-v2 predictions, standard training.
  Neighbouring-difference loss.csv    The same model trained with the neighbouring-difference loss.
  Open all three under Analyze data > Expression tables. Each prediction file becomes one model,
  named after the file. Spots and genes are matched by name, so row order does not matter.
  Optional columns: region (a pathology label for each spot) and patient_id.

score-table/
  neighbouring-pairs-scores.csv       Fine-scale q, broad-scale q and Overall Pearson r of 11 spot models
                                      in 9 HEST specimens, trained with non-neighbouring and with
                                      neighbouring pairs (the data of Figure 6a,b).
  Open under Analyze data > Computed results. One row per model, task and specimen.

analysis-bundle/
  hest-analysis.json                  A finished HEST run for TENX118 with band scores, gene scores and
                                      expression, written by `stdetail export`. Open under Computed results.
