Try your data

A counts matrix and a sample design table naming a two-level factor. Nothing on this path knows about cancer — tumor and normal are values you supply, not values the code expects.

Nothing here is uploaded. The files are read in your browser with File.slice().text() and never leave your machine: the checker has no upload, no form action, and makes no network request of any kind — your data is never part of any request this page sends. (The page itself is an ordinary page on this site and loads its web font the way the rest of them do.) It runs the same checker bindsight ui runs, from the same file.

Counts matrix (.tsv, .txt, .gz)

Genes as rows, samples as columns, integer counts.

Sample design table (.tsv, .txt, .gz)

One row per sample, with a column naming the condition.

Files are checked the moment you choose them — shape, separator, and whether the sample names in the two tables agree. The contrast levels are read from your design file and offered back to you rather than typed: a mistyped level used to surface as a generic pipeline failure minutes later, after the cohort had already been processed.

If the checks pass, the same files will run under bindsight discover.