How To Determine Whether Proteins Are Contained In Assembled Contigs

Question

[HG]

Hi everyone, I want to check whether proteins are contained in assembled contigs or not ?

For this After De-novo assembly I made a local data base of contig file, which are in fasta formate :

makeblastdb -in contigsH131.fasta -dbtype nucl -out ecolidb.db

In next step i want i tried tblastn and getting some error message like:

 ./tblastn -num_threads 2 -comp_based_stats F -query refence.fasta -db ecoli.db -out report.bls

Error msg

BLAST Database error: No alias or index file found for nucleotide database [ecoli.db] in search path [/home/hiren/Desktop/test/ncbi-blast-2.2.28+/bin::

Can anyone please help me out.

Thank you so much.

[Manu Prestat]

Have you tried to set the complete db path? It's like the program is looking it up in the blast bin directory.

[HG]

No i did not set . Because my working path in bin directory and database also present in bin. So do you think should i set complete path??

[5heikki]

If I were you, I'd first predict proteins, and then create a protein db to blast against. Anyway, in your makeblastdb command you call the db "ecolidb.db" and then you try to blast against "ecoli.db". There's your problem.

[HG]

Ok I appreciate your idea. If i am correctly understand you idea : i will predict the protein of reference genome first and make a local protein db then i will blast my contig.db against it. Please correct me if i am wrong ..

[5heikki]

Sorry, what I meant to write is that first you predict proteins from your contigs with e.g. FragGeneScan, and then you blast them against a protein db, e.g. nr or refseq_protein, or maybe some more specific db that might serve your goals better. The upside of blasting against nr or refseq_protein is that you can link your hits to KO numbers (if you have access to KEGG FTP) and then see also what modules/pathways are present..

[HG]

Trying according to your suggestion but getting some error any idea:

./run_FragGeneScan.pl -genome=./mydata/contigsH131.fasta -out=./mydata/contigsH131.test  -complete=1  -train=illumina_10
no. of seqs: 191

Use of uninitialized value $sff in split at ./post_process.pl line 141, <SEQ> line 88230.

[5heikki]

No clue, I'm using this version of FGS. If you're predicting from contigs, the command is: ./FragGeneScan -s contigs.fasta -o FGS -w 1 -t complete

[HG]

ok ...let me check and let you know

[HG]

Error is like that: any idea

./FragGeneScan -s contigsH131.fasta -o FGS -w 1 -t illumina_1

no. of seqs: 189

Segmentation fault (core dumped)

[5heikki]

Is that with the "mg-rast" version I linked, or the one from http://omics.informatics.indiana.edu/FragGeneScan because I also had segmentation errors with the latter..

[HG]

its: http://omics.informatics.indiana.edu/FragGeneScan/