Download

The training tables and the fitted models will be published here. They are not open yet.

What will be here

Three things, once they are ready to stand on their own:

• The absorption table. The 781 compounds carrying a continuous percent-absorbed label, with the published values behind each one rather than only the median, so that the disagreement described on How it works can be read directly instead of taken on trust.
• The two peptide tables. The Caco-2 arm and the passive arm, as two separate files. They will not be shipped as one file with an assay column, because a single file is an invitation to pool them and the two arms are never pooled.
• The fitted models. The descriptor sets, the scalers, the committees and the applicability thresholds, so a prediction made here can be reproduced somewhere else and checked.

Each will carry its provenance, its license and its checksum when it is posted. None of those exist yet, so none of them are printed here.

Status

Checking whether anything has been published.

This page asks the service directly rather than hard-coding an answer, so the moment a file is published it will appear here without the page being rewritten. If the service is not answering, that is reported as itself and not as an empty list.

In the meantime

Everything the models were fitted to is described on Data, including the consensus descriptors for all three models and the numbers behind every claim on this site. The method is on How it works, and the sources are on References. If you need a table before it is posted here, use Talk to us in the header and say which one.