If you can correlate the sequence to TargetDB/PDB/SwissProt sequences via BLAST search, you should be able to identify the originating nucleotide sequence. Some of the structural genomics centers provided the DNA sequence of their clones in TargetDB (now called TargetTrack). Check out NYSGXRC.xml.gz for example, which gives the nucleotide sequences even if the gene was codon-optimized and synthesized (thousands of examples). The PDB will have accession numbers, and many of those can be traced back to a DNA sequence (as can SwissProt entries). Don't reverse translate from amino acid to nucleotide as someone suggested, since you want to look at codon preferences.

Lookup NCBI with accession no

You could submit the list of accession numbers to NCBI Batch Entrez http://www.ncbi.nlm.nih.gov/sites/batchentrez

Thanks for suggestion, but access number is not available. Unfortunately tblastn (protein search against translated nucleotide database) also fail. I check reason it is because NCBI allow to search against "nt" (non-redundant nucleotide sequence) which does not have all nucleotide sequences.

What is reported in the title of the fasta file of your protein? I find very hard to believe no protein returned a blast result against the nr database. Even a match with a portion of the protein is expected...... Let us know!

If you want you can send me a private message with some example of proteins that did not match any nucleotide sequence......

If you can correlate the sequence to TargetDB/PDB/SwissProt sequences via BLAST search, you should be able to identify the originating nucleotide sequence. Some of the structural genomics centers provided the DNA sequence of their clones in TargetDB (now called TargetTrack). Check out NYSGXRC.xml.gz for example, which gives the nucleotide sequences even if the gene was codon-optimized and synthesized (thousands of examples). The PDB will have accession numbers, and many of those can be traced back to a DNA sequence (as can SwissProt entries). Don't reverse translate from amino acid to nucleotide as someone suggested, since you want to look at codon preferences.

Thanks for advise and support, following are datasets used in previous methods SolPro and Proso II for which we wish to extract corresponding nucleotides.

http://download.igb.uci.edu/SOLP.fa

Proso II server (suppl) http://mips.helmholtz-muenchen.de/prosoII/prosoII.seam

http://mips.helmholtz-muenchen.de/prosoII/img/Suppl_files.zip

These files have protein in fast format, @Sauder, thanks for suggestion, we are working on it. Unfortunately, at http://sbkb.org/tt/downloads.html , dnaTagetSeqs does not contain DNA corresponding to all protein target and rnaTargetSeqs is empty. We are exploring possibility to extract DNA from targets.xml (reference datbase). So far we are able to extract around 1000 DNA sequences corresponding to proteins in SolPro.

We are interested to understand why researchers unable to make all protein soluble (proper folding), though high expression of gene (in E. Coli) is not a issue (replace rare codon with major codons).

Gajendra are you sure NCBI produced no results? I tried to run a tblastn with the first sequence which I found at http://download.igb.uci.edu/SOLP.fa and found several matches.....

@Caminolo, yes we got several matches no matches give 100%. Problem is out of several matches which is original/actual DNA sequence corresponding to query protein sequence. Here, we can not afford ambiguity, we need original DNA sequence responsible for protein.

Please search homology protein of your interested protein and then try to find out nucleotide sequence

Vinod thanks for suggestion but this is not a simple problem.

Do you know the organism the proteins belong to? The first sequence I submitted to NCBI returned several perfect matches. E.g.:

Mycoplasma pneumoniae gyrase B (gyrB) gene, partial cds, DnaJ homolog, DnaN (dnaN), Soj (soj), DnaA (dnaA), and CysA homolog genes, complete cds

Sequence ID: gb|U34816.1|MPU34816Length: 14177Number of Matches: 1

@GR:

I feel you should try SciFinder/CAS for exact sequences of the desired proteins. As CAS has got a large number of protein sequence data-sets also in the CAS database. This way you will get exact result with accession number, protein structure Registry number, which you can exploit in other databases. Actually, You have to pay for accuracy. I know you can.

Actually, I had used SciFinder for getting sequence as well as protrin function information for insilico prediction of medicinal properties of the medicinal plants and marine biota). I worked. I presented a paper also in your conference as poster. I am confident that you will get what you want.

Regards

rks

Thanks Dr Sharma for your excellent suggestion, unfortunately I do not use commercial software or services in my research. Our research is based on open source and our output also falls in open source.

For sequences in TargetDB/Target Track, the structural genomics centers are expected to provide their plasmids to the DNASU material repository at Arizona State University. You could contact the scientific liaison there to see if there's an easy way to obtain all these sequences to do your sequence comparison; if you find exact or near exact matches (be aware of tags at the terminii, etc), then DNASU should be able to provide accession numbers and other information associated with that plasmid.