Changes between Version 32 and Version 33 of NIAID2022

Timestamp:

Dec 8, 2022, 1:54:00 PM (3 years ago)

Author:

Elaine Meng

Comment:

--

NIAID2022

@@ -59 +59 @@
 = Meeting Minutes =
 
+== 12/8/2022 ==
+
+Phil, Kristen, Michal, Meghan; Eric, Elaine, TomG, Scooter
+
+Kristen made tickets for several pipeline-stopping small-molecule CIF structures: was many before, but now only 10.  If fixes won't be soon enough we can special-case the 10 out for now. Eric: Greg was in the middle of working on small-molecule CIFs. Some work, but some others may not.  Eric: one current problem is an error in falling back to small-molecule CIF when the mmCIF reader falls. Unfortunately this is Greg's area and he may not be able to finish the fixes soon enough.  Phil: we can special-case the 10 problem structures for now. Kristen: going forward, need to track the errors so that they can get fixed later.  Plan for beta release is a restricted set of users for about 2 months (until approximately the end of February).
+
+Another problem is "SDF: list index out of range".  Eric: these files have a different type of line returns in the charge section (just carriage returns without newline).  I can fix this specific error by looking for embedded carriage returns and doing an extra line-splitting step.  But will other bad-format errors arise?  Can't foresee them all... Kristen: we can raise a format error. Scooter: should reach out and find out if a commonly used program is generating this weird type of file.  Kristen: OK we will reach out and get back to you.
+
+Another problem: ValueError: all hydrogen names for /A NH2 37N taken. Eric: the NH2 capping residue is disconnected so it gets an extra hydrogen. Unfortunately it may be a Greg issue if the mmCIF reader is omitting a bond. Phil: the mmCIF file (biounit version from PDBe) just doesn't have the bond info, but Pymol adds it.  Eric: maybe based on distance. #8139.  Elaine and TomG: there are two biounit fetch sources, rcsb_bio and pdbe_bio and they are done differently.  One uses multiple models and the other puts them all into one model with more chains. Eric: maybe script should just remove and re-add all bonds. It won't take long. Tried on example 2mag: ~bond #1; bond #1
+
+Linux build broken since 12/6. TomG will try to fix today and comment on the ticket.
+
+Another: NMR structure of DNA treated as small molecule. Eric: caused by assembly file again. TomG: this file (3dpx) is weird, has multiple sections for a single model. An endmdl record is reached when only a subset of the atoms for a model has been read.
+
+Another: need to decide on why the 500-atom criterion is in the ribbon preset, adjust it or remove it. Phil: couldn't come up with the rationale, maybe we should just remove it. TomG: or change behavior of "struts" wrt atom display, but probably better to remove the criterion first and see if that gives what you want.  Phil: let's try that.
+
 == 11/10/2022 ==