Which BLAST should I use?

[SDPA_Pet]

Hello, I have some sequencing data (soils) from one functional gene (nitrogen fixation gene nifH). I was wondering which BLAST should I use. I just want to use blast find out the taxonomic info of these data.

When I did my WGS metagenomics, I use BLASTX against NCBI nr database, because I don't which gene in my metagenomes.

For this single gene data? Should I use BLASTn or BLASTX? I think I should use BLASTn right? or I still use BLASTX?

Also which database should I choose? nt or nr? What is the difference between these two databases? Is nt database more comprehensive? or nr is more focused on protein / functional gene?

Someone told me nr database is a reduced version of nt. I should use nt or better results.

[GenoMax]

nt: Nucleotide sequence database contains entries from all traditional divisions of GenBank, EMBL and DDBJ. Sequences from bulk gss, sts, pat, est, htg divisions plus wgs and env_nt entries are excluded.

nr: Non-redundant protein sequences with entries from GenPept, Swissprot, PIR, PRF, PDB and RefSeq

You could get the sequences from one of the searches below:

https://www.ncbi.nlm.nih.gov/gene/?term=nifH[sym]

https://www.ncbi.nlm.nih.gov/gene/?term=nifH

https://www.ncbi.nlm.nih.gov/protein/?term=nifH

[SDPA_Pet]

You could get the sequences from one of the searches below:

https://www.ncbi.nlm.nih.gov/gene/?term=nifH[sym]

https://www.ncbi.nlm.nih.gov/gene/?term=nifH

https://www.ncbi.nlm.nih.gov/protein/?term=nifH

Hi GenoMax,

It is a large dataset and I will use offline BLAST. That's why I need to choose the program.

[GenoMax]

If you only get the nifH gene sequence (DNA/protein from those three links above) then your database will become much more manageable. You may want to try both blastn and blastx against this reduced dataset first.

If that is not satisfactory then you may need to start looking at nt/nr. But then you would have a lot of alignments to weed through.

How confident are you about the ability of pulling down just nifH with the primers that you designed? Is the data from sanger/fasta files?

[SDPA_Pet]

If you only get the nifH gene sequence (DNA/protein from those three links above) then your database will become much more manageable. You may want to try both blastn and blastx against this reduced dataset first.

If that is not satisfactory then you may need to start looking at nt/nr. But then you would have a lot of alignments to weed through.

How confident are you about the ability of pulling down just nifH with the primers that you designed? Is the data from sanger/fasta files?

The primers have been designed and used for many years and should be fine. So, suppose there is no problem with primer design.

Should I just choose BLASTx? What's the difference between BLASTn and BLASTx in my special case. I don't think it is much difference in my case? Supppose my primers are acurate, I get all sequences coding for proteins. I should try BLASTX first, right? This is what you mean, right?

[GenoMax]

If you primers are accurate (and since this is bacterial data so no worries about splicing) then either should work.

Again if you start with the reduced datasets (with just nifH sequences DNA or protein) then you can test both blastn/blastx quickly. Make sure you get the sequence titles/use parse_seq_id option when you make the blast databases.