1ffk missing nucleotides?

[erikbike]

Hi Xiang-Jun,

I am comparing the structures output by DSSR to those in FRABASE.  For the structure 1ffk, the DSSR output contains something like:

U&A
(&)

The structure in FRABASE for this same regions shows:

UCUAA
(---)

http://rnafrabase.cs.put.poznan.pl/?act=pdbdetails&id=1FFK

Note the CUA that is missing from DSSR's output.  It appears that DSSR is dropping the nucleotides between chains.  Therefore the whole chain output is incomplete if ampersands (&) are removed.

Is there a way to have DSSR output the whole sequence with missing nucleotides separated by "-" characters, as in the output from FRABASE?

Thanks,
Erik

[xiangjun]

Hi Eric,

I am aware of the missing sequence info in DSSR-derived dbn -- it is on purpose. Internally, DSSR checks only ATOM/HETATM records in a PDB or PDBx/mmCIF input file. Specifically, the program does not care about the SEQRES records. Presumably it should be straightforward to take SEQRES into consideration. However, the program then has to consider each 3-letter nucleotide id to its 1-letter shorthand mapping, and check for consistency between sequence and atomic coordinates. Another practical consideration is that many PDB files outside wwPDB do not have SEQRES records at all. So it is a deliberate decision to leave the sequence information out for those structures that have disordered region (thus without corresponding atomic coordinates).

You may well want to use FRABASE for your project if the missing sequence info is a concern. Alternatively, you may also want to try RNApdbee from the same group.

Best regards,

-- Xiang-Jun