You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We detected extremely large computational times in the Side Chain Prediction algorithm when using Impact templates generated with peleffy. The reason of this undesired behaviour seems to be a consequence of a bad atom ordering in these templates. Atoms are not properly sorted by their parent indices.
Solution
Fix the ordering in the Impact template.
The text was updated successfully, but these errors were encountered:
Now we know that the atoms in the Impact template need to be sorted by two criteria:
Atom location: first atoms from the core, then atoms from the side chain
Parent index: ascending order of parent indices according to the molecular graph
The problem was that, although atoms where correctly sorted by atom location, they were not following the right criteria with parent indices. As a result, PELE was able to load molecular topologies with no problem but, in some cases, this bad ordering lead to problems when running the side chain prediction algorithm. In these cases, side chain prediction algorithm could require a large computational time (e.g more than 1000s).
This issue seems fixed with the correct order of the atoms in the Impact template file.
Description
We detected extremely large computational times in the Side Chain Prediction algorithm when using Impact templates generated with peleffy. The reason of this undesired behaviour seems to be a consequence of a bad atom ordering in these templates. Atoms are not properly sorted by their parent indices.
Solution
Fix the ordering in the Impact template.
The text was updated successfully, but these errors were encountered: