3DNA Forum

Questions and answers => RNA structures (DSSR) => Topic started by: febos on January 24, 2016, 12:45:05 pm

Netiquette · Download · News · Gallery · Homepage · DSSR Manual · G-quadruplexes · DSSR-Jmol · DSSR-PyMOL · DSSR Licensing · Video Overview· RNA Covers

Title: non-one-char chain identifiers
Post by: febos on January 24, 2016, 12:45:05 pm
Good afternoon, Dr. Lu

I've noticed an issue with long chain identifiers in the section with dot-brackets.

Example:
Entry - 1VY6

fragment from DSSR output header:
Code: [Select]
no. of DNA/RNA chains: 12 [AA=1498,AV=13,AW=2,AX=76,BA=2819,BB=120,CA=1503,CV=12,CW=2,CX=76,DA=2800,DB=120]fragments from dbn section:
Code: [Select]
>1vy6-1-A #1 nts=1498 [chain] RNA*
>1vy6-1-A #2 nts=13 [chain] RNA
>1vy6-1-A #3 nts=2 [chain] RNA
>1vy6-1-A #4 nts=76 [chain] RNA
>1vy6-1-B #5 nts=2819 [chain] RNA*
>1vy6-1-B #6 nts=120 [chain] RNA
>1vy6-1-C #7 nts=1503 [chain] RNA*
>1vy6-1-C #8 nts=12 [chain] RNA
>1vy6-1-C #9 nts=2 [chain] RNA
>1vy6-1-C #10 nts=76 [chain] RNA
>1vy6-1-D #11 nts=2800 [chain] RNA*
>1vy6-1-D #12 nts=120 [chain] RNA

I cannot recover one-to-one correspondence from these lines. I mean, of course in the given case i am able to find correspondence not from identifiers but from lengths, but this is just a lucky example.
Could you fix this, please?


Best regards,
Eugene
Title: Re: non-one-char chain identifiers
Post by: xiangjun on January 24, 2016, 02:15:12 pm
Hi Eugene,

Thanks for reporting this inconsistency in dbn output with chain names longer than 1-char. The current setting reports only the first letter, which can cause problems, as you see here.

Would changing the current

Code: [Select]
>1vy6-1-A #1 nts=1498 [chain] RNA*
>1vy6-1-A #2 nts=13 [chain] RNA
>1vy6-1-A #3 nts=2 [chain] RNA
>1vy6-1-A #4 nts=76 [chain] RNA

to

Code: [Select]
>1vy6-1-AA #1 nts=1498 [chain] RNA*
>1vy6-1-AV #2 nts=13 [chain] RNA
>1vy6-1-AW #3 nts=2 [chain] RNA
>1vy6-1-AX #4 nts=76 [chain] RNA

fits the bill? Please confirm.

I will release an updated DSSR soon.

Xiang-Jun
Title: Re: non-one-char chain identifiers
Post by: xiangjun on January 24, 2016, 02:46:33 pm
Hi Eugene,

DSSR has been updated to v1.4.9-2016jan25 which should have fixed the inconsistency in chain names for the dbn section, as tested on 1vy6 (http://www.rcsb.org/pdb/explore/explore.do?structureId=1vy6).

Please check the new DSSR release (still shown as v1.4.8-2016jan16 on the download page -- I'll update the release note later). Please report back how it goes.

Best regards,

Xiang-Jun


Xiang-Jun
Title: Re: non-one-char chain identifiers
Post by: febos on January 25, 2016, 08:13:10 am
Good afternoon, Dr. Lu

Sorry for late reply.
This fix should be great.

Just in case:
In pdb entries I saw chain identifiers up to 4 chars (I do not know if it can be longer or not, but 4 is the largest I have seen).
Does the last update of DSSR take it into consideration?

Best regards,
Eugene
Title: Re: non-one-char chain identifiers
Post by: xiangjun on January 25, 2016, 09:12:00 am
Hi Eugene,

Yes, 4-char chain names should have been taken care of by DSSR. As far as I know, in mmCIF formatted structures, chain names can have up to 4 characters.

Please do not hesitate to report back any bug/inconveniency you may find in DSSR.

Best regards,

Xiang-Jun

Created and maintained by Dr. Xiang-Jun Lu [律祥俊] (xiangjun@x3dna.org)
The Bussemaker Laboratory at the Department of Biological Sciences, Columbia University.