Foci Lab

Screen idr0010 · your browser does the work · exploratory

IDR plate mode

Choose one plate of the genome-wide 53BP1 screen (Doil et al. 2009). Your browser downloads that plate's wells from our cache, counts nuclei and foci with the Foci Lab engine, and lines the result up with the authors' own per-well numbers. Our server never runs the analysis; it only hands out images.

Pairing of images and annotation rows: inferred, unconfirmed.

The image grid of a plate and the authors' table number the wells differently. The default here is column-major (inferred): the annotation row with Well Number N goes with the image at row (N−1) mod 12, column (N−1) div 12, i.e. the image grid is counted column by column (A1, B1, … L1, A2, …). The alternative, by IDR name (image A2 with row A2), is how the site was first built. The column-major rule fits the authors' own cell counts far better (plate 7-14: nuclei vs cells Spearman 0.56 against 0.07; foci per cell 0.30 against −0.01), and so do lethal controls, but that is indirect evidence: it is empirically inferred and unconfirmed by IDR or the authors. Both pairings stay selectable and are shown side by side in the table and the comparison cards. See docs/EXPERIMENT_PROTOCOL.md §15.4.

Read this first.
  • The counter is exploratory and not validated against manual counts. A low or high correlation with the authors' numbers does not by itself say which side is wrong.
  • 53BP1 foci mark DNA double-strand breaks anywhere in the genome; they are not telomere-specific.
  • Images come from the public Image Data Resource (idr0010, CC BY 4.0) through a cache on this site; an image is fetched from IDR once and then served from the cache. Nothing you do here uploads anything.

1. Pick a plate

Engine checking…
Engine parameters (optional)

Defaults come from the engine. To change some, enter a JSON object such as {"min_height": 14}; it is merged over the defaults and recorded in the provenance export.