Hi Zhang Lab team!
I used the ThreaDomEx server on many proteins and I realized this morning that the files jobID.result, jobID.sd and jobID.score didn't reflect the result showed on the server output page. I did not inspect every jobID.dcd file so I do not know if those files have the same issue.
For example, job ID T00235e7d7b9 show 4 domains on the web server but on the T00235e7d7b9.result file, there is 5 domains. It seems that residues 693 and 694 have scores of 0.556 and 0.558, which is below the cutoff of 0.56. This should separates the domains but on the web page, it didn't.
Also, job ID T002418a7aac result file show 3 domains instead of the 4 domains showed on the web page. The domain 4 beginning at residue 400 (showed on the web server output page) have a score below the cutoff (score of 0.326, cutoff of 0.56). This isn't showed on the result file.
Another example is job T002307667e4 where the scores aren't the same on the web and in the file. This results in a 2 domain protein on the server and in a 1 domain protein according to the downloaded files. For example, residue 104 have a score of 0.687 in T002307667e4.result file and 0.75578 on the web page. This is the case for almost all (if not every) residue.
Finally, some jobs have domains predicted at different positions when comparing the result file and the web page. As they are supposed to have the same information, there is an issue with those jobs too. (Example: T00231b30e5b)
there is 13 of the 20+ jobs I sent that have similar issues. Here is the list of the other jobs with issues if it can help:
T002345147b0
T002296c6345
T002259d434d
T00189de0722
T00191bf124b
T0023776cf23
T0024472747b
T0024397f2fe
T00239fad1f9
Thank you
Vincent