P. generosa functional gene annotations

[ocattau]

After I have my count data for a single library for P. generosa, I need to annotate my table with gene functions. Do we already have a Blast for v5 of the transcriptome? If so where would I find it?

[kubu4]

On Mox: /gscratch/srlab/blastdbs/ncbi-sp-v5_20210224/

[ocattau]

Okay new question for Mox newbs, how to I download this file in a way I can use R to analyze it?

[kubu4]

Which file?

[ocattau]

ncbi-sp-v5_20210224/swissprot.tar.gz

I know I need to use rsync but where can I upload it to? Does it have to be gannet?

[kubu4]

You don't want/need that file. You'll want to run BLAST on Mox, which would utilize that BLAST database.

[ocattau]

Ah That is true, okay so there isn't full transciptome in Mox then. I have to blast again.

[kubu4]

You'll want to rsync your query file (e. g. a FastA file) to your directory on /gscratch/srlab/ (you might need to create your directory) or /gscratch/ scrubbed/.

Then, run BLAST and use that BLAST database listed above.

[kubu4]

The video you sent uses the fasta so I got confused

Sorry, that's an unfortunate/confusing naming thing. If you look in that directory, you'll see what the files are called. Here's an example:

ls -l /gscratch/srlab/blastdbs/UniProtKB_20190109
total 574976
-rw-r--r-- 1 samwhite hyak-srlab       263 Jan  9  2019 readme.txt
-rw-r--r-- 1 samwhite hyak-srlab 275880518 Jan  9  2019 uniprot_sprot.fasta
-rw-r--r-- 1 samwhite hyak-srlab 106931528 Jan  9  2019 uniprot_sprot.fasta.phr
-rw-r--r-- 1 samwhite hyak-srlab   4471264 Jan  9  2019 uniprot_sprot.fasta.pin
-rw-r--r-- 1 samwhite hyak-srlab 201309527 Jan  9  2019 uniprot_sprot.fasta.psq

The -db argument requires you to specify the basename of the database. So, in this instance, you would specify uniprot_sprot.fasta. Although there is an actual FastA file (this is all very easy to follow, isn't it? Ugh.), the -db argument is looking for the three files ending in .phr, .pin and .psq. So, it expects the basename for all three of those files, which happens to be: uniprot_sprot.fasta.

Related to this, you should try to use one of the BLAST DBs we have available. I'd recommend against moving/copying something to your own directory. The primary reason for this recommendation is to improve reproducibility. If you use a BLAST DB that's established in one of the shared locations, it will have documentation about when/how/who generated it and other people can use the same DB to help test out reproducibility. Plus, using one of the established DBs will help us troubleshoot if things go wrong.

EDITED: Acknowledge confusion of existence of a FastA file in the BLAST DB directory example above.

[sr320]

it will have documentation about when/how/who generated it

we just looked at the folder and did not find a readme, so it was not clear to us what all the databases were.

[kubu4]

Which folder? I'm not even sure what database is trying to be utilized.

[sr320]

I believe we were looking at '/gscratch/srlab/blastdbs'

[kubu4]

There are a bunch of DBs directories in there. There are readmes present in each of the BLAST DBs which describe how/who/when/etc.

E.g.

cat /gscratch/srlab/blastdbs/uniprot_sprot_20200123/readme.txt


Directory containing files for BLASTing against the Uniprot SwissProt database.

---

FILES

- uniprot_sprot.fasta: Uniprot SwissProt protein FastA file. Downloaded 20200123.

- uniprot_sprot.dmnd: DIAMOND Uniprot (Reviewed) BLAST database generated with the following command:

makedb --in uniprot_sprot.fasta --threads 27 --db uniprot_sprot --taxonmap ../ncbi-nr-20190925/prot.accession2taxid --taxonnodes ../ncbi-nr-20190925/nodes.dmp

[ocattau]

Okay starting now thread in this discussion but if I want to blast a scaffold Panopea-generosa-v1.0.fa against the geoduck db online at https://owl.fish.washington.edu/halfshell/genomic-databank/Pgenerosa_transcriptome_v5.fasta that I dowloaded from the command line into my folder in Mox, so my code looks like this:

/gscratch/srlab/programs/ncbi-blast-2.8.1+/bin/blastx \
-query /gscratch/srlab/ocattau/p.generosa/Panopea-generosa-v1.0.fa \
-db /gscratch/srlab/ocattau/p.generosa/Pgenerosa_transcriptome_v5.fasta 
-out /gscratch/srlab/ocattau/p.generosa/Panopea-generosa-uniprot_blastx.tab \
-evalue 1E-20 \
-num_threads 30 \
-max_target_seqs 1 \
-max_hsps 1 \
-outfmt 6

Again I am getting the same error message that -db /gscratch/srlab/ocattau/p.generosa/Pgenerosa_transcriptome_v5.fasta is not a database or index.

# the code for Pgenerosa_transcriptome_v5.fasta 
TRINITY_DN88518_c0_g1_i1 len=571 path=[1:0-44 2:45-525 4:526-570]
CAATAACAGCTTATATTTGGAAATGAACAGATTAACATAAGCTGGATGGGAGTCCATTGATTGGGGTGTTGATTTCTCATGTGAAAGAAATGGAATTGAGACCAATTCATAGCTTAGTTTACAAGTTTAGGGTCTGGGGTTATAAAACTTGGTTTGTAATTGATAAAATTTTAGCCATACTTGTTTGTAGTTGATAAAGTTACAGTCCTTGTAGTTATGGTAACCATATCCTCTGTATAAATAGACATGGTAAATGTTACATTATATCTGACTGTTGTTATAGTGGTGTCAGCAGGTCGCAGAAAGAAGGTAAGCTTGGCTGTGATTTGCTTTAGAAAAAATAATTAAGTGTGATTGATTGATTGATCTATTTATTGATTGGATTTGATTCTGATGAATTTCCATAAAAAAAATAAACTTGGAGTAGTAGTGTTTTGAGTAATAAATATTTTACATTTTTTCAACTTCAGTTCTGTTATCTATTGTTTTTCCTTTTTTTTCTGTAAATAATTATATACTTAGATTACGACTTTGTGAATTCAAATTTATTTAAATAATTGTGAATCAGCAC

Is this not a database?

[kubu4]

Is this not a database?

It is not. It is a FastA file.

If you want to BLAST against that, you need to create a BLAST database using the makeblastdb command (it's part of the BLAST suite of programs).

[sr320]

more importantly you DO NOT want to blast a scaffold Panopea-generosa-v1.0.fa against the geoduck db

[ocattau]

So we are back to -db=swiss-prot? When I ran the code yesterday it timed out after 12 hours so I feel like that was incorrect also. What db should I be using then?

[kubu4]

it timed out after 12 hours so I feel like that was incorrect also

Why do you think so?

BLAST-ing will take a long time.

Can you please post your SLURM script which was running successfully and timed out?