metapredict is a deep-learning consensus predictor of intrinsic disorder. Paste a sequence below, then drag across the plot to pull any region straight onto your clipboard.
These control how the continuous disorder profile is carved into discrete IDRs. Leave the threshold blank to use whatever the selected network was calibrated with (V1 = 0.42, V2 and V3 = 0.5). Changing any of these re-runs the prediction.
Drag across the plot to select a region — it lands on your clipboard. Click the red bar under the axis to grab a whole IDR. Double-click or press Esc to clear.
Rg and Re are ALBATROSS predictions of the ensemble-average radius of gyration and end-to-end distance, alongside their AFRC-normalised values (<1 means more compact than an ideal chain, >1 more expanded).
metapredict.net handles a single sequence per request. For anything bigger you have two good options:
V3 is the default and the one to use for new work — it is our most accurate predictor (See Lotthammer et al. bioRxiv). V2 and V1 are kept available so older analyses can be reproduced exactly; V1 in particular uses a different disorder threshold (0.42) and will give noticeably different IDR boundaries.
The selection panel reports ALBATROSS predictions for the region you have highlighted. If you want the full set of ensemble properties, head to the ALBATROSS Colab notebook.
Cite the ALBATROSS paper if you use the predicted dimensions in the selection panel.