Dr. Gang Chen notified me on July 13th that some LUSC patients had their vital status listed incorrectly. I confirmed the problem and immediately opened an issue on the GitHub repository: https://github.com/OmnesRes/onco_lnc/issues/1
I also immediately posted the issue in bold on the OncoLnc home page, tweeted about the issue, and pinned the tweet to my profile page.
Due to inconsistencies in the clinical files provided by TCGA, some patients had their vital status listed as 'Alive' when they should have been listed as 'Dead'. This affected the following cancers: CESC, COAD, GBM, HNSC, KIRC, LGG, LUAD, LUSC, OV, READ, SKCM, and STAD. For the majority of these cancers only around 1% of patients were affected, but GBM and OV were exceptions, and so was LUSC to a lesser extent.
Although the data present in OncoLnc was not necessarily wrong, it was not as complete as it could have been. And since OncoLnc is the only comprehensive survival analysis tool available for TCGA data the data needs to be as accurate as possible. So I quickly changed the necessary code and reran the affected Cox regressions. This affected over 100 files in the GitHub repository. I pushed the fix on July 15th.
I then updated all affected excel files present at http://www.oncolnc.org/download/ and rebuilt the 1.3 GB SQLite3 database.
As a result, the data present on OncoLnc may differ from that provided as Tables and Supplemental Tables in this article. To ensure the most accurate data you should always download directly from http://www.oncolnc.org/.