[upendra_35]

Blast2GO is very particular about the format of the sequence ID used in FASTA files when you are creating custom BLAST databases. See this webpage for details about formatting your own BLAST DB for use with Blast2GO.

A second thing to consider is do the IDs in you custom BLAST DB ('myid' in the example on the webpage) match IDs in the Blast2GO database. If the IDs don't match, Blast2GO won't be able to map them. Even if you reformatted the TAIR FASTA like the example in the tutorial and put the AT number in the proper position I'm not sure that TAIR AT numbers are in the Blast2GO database.

A workaround is to use this file from the TAIR site: At_GB_refseq_prot.gz. It is the same protein set but uses the NCBI RefSeq IDs and GI numbers for the IDs. For example:

>gi|240256448|ref|NP_200529.4| PSD2 (phosphatidylserine decarboxylase 2); phosphatidylserine decarboxylase [Arabidopsis thaliana]

Format this file according to the tutorial. You are sacrificing the AT numbers but it's pretty much a guarantee that the NCBI GIs are in the Blast2GO database.

Thank you very much kmcarr, very useful tips. Can't wait to start my Blast2Go annotation now. Thanks again.

[kmcarr]

I've tried changing format but it doesn't work. I guess we should use TAIR10_pep... for BLAST DBs which is protein files.

Blast2GO is very particular about the format of the sequence ID used in FASTA files when you are creating custom BLAST databases. See this webpage for details about formatting your own BLAST DB for use with Blast2GO.

A second thing to consider is do the IDs in you custom BLAST DB ('myid' in the example on the webpage) match IDs in the Blast2GO database. If the IDs don't match, Blast2GO won't be able to map them. Even if you reformatted the TAIR FASTA like the example in the tutorial and put the AT number in the proper position I'm not sure that TAIR AT numbers are in the Blast2GO database.

A workaround is to use this file from the TAIR site: At_GB_refseq_prot.gz. It is the same protein set but uses the NCBI RefSeq IDs and GI numbers for the IDs. For example:

>gi|240256448|ref|NP_200529.4| PSD2 (phosphatidylserine decarboxylase 2); phosphatidylserine decarboxylase [Arabidopsis thaliana]

Format this file according to the tutorial. You are sacrificing the AT numbers but it's pretty much a guarantee that the NCBI GIs are in the Blast2GO database.

[upendra_35]

I've tried changing format but it doesn't work. I guess we should use TAIR10_pep... for BLAST DBs which is protein files.

Sorry my bad.....i have used pep file and not cDNA file.

[hugh_hang]

I got the same problem. When i tried to import the xml file that was blatsted against Arabidopsis TAIR10 cDNA reference, the mapping and annotation steps failed in blast2go software. So i guess it is something to do with the format of xml file generated using "nr" database and "TAIR10" database.
Does anybody know a way to modify this xml file so that i can import to blast2go?

Thanks
Upendra

I've tried changing format but it doesn't work. I guess we should use TAIR10_pep... for BLAST DBs which is protein files.

[upendra_35]

I got the same problem. When i tried to import the xml file that was blatsted against Arabidopsis TAIR10 cDNA reference, the mapping and annotation steps failed in blast2go software. So i guess it is something to do with the format of xml file generated using "nr" database and "TAIR10" database.
Does anybody know a way to modify this xml file so that i can import to blast2go?

Thanks
Upendra

[hugh_hang]

What made you think that there were improper tabs or returns that needed removing?

Please post an example of the BLAST output before you edited it (just a couple of dozen lines is enough).

********************
<BlastOutput_param>
<Parameters>
<Parameters_matrix>BLOSUM62</Parameters_matrix>
<Parameters_expect>1e-06</Parameters_expect>
<Parameters_gap-open>11</Parameters_gap-open>
<Parameters_gap-extend>1</Parameters_gap-extend>
<Parameters_filter>L;</Parameters_filter>
</Parameters>
</BlastOutput_param>
********************
this is what blast2go formatted and that below is what local blast formatted.
********************
<BlastOutput_param>
____<Parameters>
________<Parameters_expect>1e-06</Parameters_expect>
________<Parameters_gap-open>11</Parameters_gap-open>
________<Parameters_gap-extend>1</Parameters_gap-extend>
________<Parameters_filter>L;</Parameters_filter>
____</Parameters>
</BlastOutput_param>
********************
("_" stands for space)
additionally, blast2go uses newline(\n) to switch line and my local blast program in windows uses return & newline(\r\n), which I suspect to impact.

[kmcarr]

I cleared the unproper tabs(\t) & enters(\r\n) and it still doesn't work. I have to doubt if it's because there are too fewer sequences that have found hits or if I should use protain DBs instead of nucleotide ones.

What made you think that there were improper tabs or returns that needed removing?

Please post an example of the BLAST output before you edited it (just a couple of dozen lines is enough).

[hugh_hang]

Are you sure the output from the BLAST search is properly formatted and contains valid hits?

I cleared the unproper tabs(\t) & enters(\r\n) and it still doesn't work. I have to doubt if it's because there are too fewer sequences that have found hits or if I should use protain DBs instead of nucleotide ones.

[hugh_hang]

cDNA-AFLP you say? O.K.

"pep" = peptide, meaning that these FASTA files contain the protein (amino acid) sequence translated from the predicted CDS.

Given the method used to isolate your material for sequencing can be assumed that the sequences are derived from protein coding genes. Amino acid sequence is more conserved than the underlying nucleic acid sequence so comparing across species using amino acid sequences is more sensitive than comparisons based on DNA sequence. This is why I suggested using the Arabidopsis protein (TAIR10_pep_*) database as a target in a BLASTX search with your cDNA query sequences. BLASTX will translate your query sequences in all 6 possible reading frames and compare those amino acid sequences for similarity.

I see. Thank you.

[kmcarr]

Sir, I've done what you recommanded me to do. But after I imported the XML file, the mapping and annotation seems not recognise the blast result and give 0 sequences mapped or annotated, my database id from TAIR10_blast_set which are nucleotide queries fasta files.

Are you sure the output from the BLAST search is properly formatted and contains valid hits?

[kmcarr]

My sequences are cDNA-AFLP results, does that mean "TAIR10_cdna_..." is a better option? By the way, what does the "pep" in "TAIR10_pep_..." mean?

cDNA-AFLP you say? O.K.

"pep" = peptide, meaning that these FASTA files contain the protein (amino acid) sequence translated from the predicted CDS.

Given the method used to isolate your material for sequencing can be assumed that the sequences are derived from protein coding genes. Amino acid sequence is more conserved than the underlying nucleic acid sequence so comparing across species using amino acid sequences is more sensitive than comparisons based on DNA sequence. This is why I suggested using the Arabidopsis protein (TAIR10_pep_*) database as a target in a BLASTX search with your cDNA query sequences. BLASTX will translate your query sequences in all 6 possible reading frames and compare those amino acid sequences for similarity.

[hugh_hang]

The default configuration for online BLAST through the Blast2GO GUI application uses the NCBI QBlast service which provides only those databases listed; you can not add custom databases to this method.

Your alternatives to use different databases are:

1) Set up your own WWW-BLAST service (or find access to someone's who will share) which has or can be customized with the databases you want. Edit the blast2go.properties file on your local computer to designate this WWW-BLAST server as the default source for running your online BLAST searches through the Blast2GO GUI.

2) Run your BLAST search using a standalone (command line) BLAST installation against your custom database. Be sure to configure your BLAST search to output the results in XML format. Launch Blast2GO and load your FASTA sequence file as normal. From the File menu select "Import->Import Blast Results". Select your XML file (or files) for import. Once the BLAST results have been imported proceed with Mapping and Annotation as usual.

I recommend option #2 because it is easier and more scalable.

Sir, I've done what you recommanded me to do. But after I imported the XML file, the mapping and annotation seems not recognise the blast result and give 0 sequences mapped or annotated, my database id from TAIR10_blast_set which are nucleotide queries fasta files.

[hugh_hang]

Just click on the link chadn provided in his reply. It will take you directly to the correct FTP directory with various FASTA files. Read the Readme_blastdatasets_TAIR10.txt file for a description of what each one is.

I, myself would choose the "TAIR10_pep_20110103_representative_gene_model_updated" (or "TAIR10_pep_20101214_updated") and assuming you are BLASTing with nucleotide queries run BLASTX against this protein dataset.

My sequences are cDNA-AFLP results, does that mean "TAIR10_cdna_..." is a better option? By the way, what does the "pep" in "TAIR10_pep_..." mean?

[kmcarr]

Originally Posted by chadn737
Here is the ftp site for the TAIR10 blast sets. You probably want one of the cds or cDNA files:

ftp://ftp.arabidopsis.org/home/tair/...R10_blastsets/

Thank you, I tried clicking "download" in the TAIR website, but I'm really confused by its dendroid file structure.

Just click on the link chadn provided in his reply. It will take you directly to the correct FTP directory with various FASTA files. Read the Readme_blastdatasets_TAIR10.txt file for a description of what each one is.

I, myself would choose the "TAIR10_pep_20110103_representative_gene_model_updated" (or "TAIR10_pep_20101214_updated") and assuming you are BLASTing with nucleotide queries run BLASTX against this protein dataset.

[hugh_hang]

Here is the ftp site for the TAIR10 blast sets. You probably want one of the cds or cDNA files:

ftp://ftp.arabidopsis.org/home/tair/...R10_blastsets/

Thank you, I tried clicking "download" in the TAIR website, but I'm really confused by its dendroid file structure.