BLASTX Issues

[salamay]

I am attempting to run blastx from the command line, and it seems to be very slow (~5-6 minutes for 10 100bp sequences). I compared this to blastx on the NCBI website, which took only 30 seconds. This indicates that there is an issue with the command line usage of blastx, but I don't know where to start examining this issue.

I made sure that the db used in the command line and on the ncbi website was the same (nr). I am using all default parameters.

Any help would be appreciated.

Thanks

[GenoMax]

That is an apples to oranges comparison.

Hardware NCBI uses for Blast searches is going to be tremendously more powerful than anything you are using locally.

What kind of hardware are you running your searches on? How much RAM are you allocating to your searches locally?

[salamay]

The machine I run the searches on is a 48 core 2.7ghz with 264 Gb RAM, but not sure how much of that is allocated for the search (nor do I know how to specify allocation). I run the search with default parameters so num_threads is set to 1 by default. Even though, I don't think it should be that slow. Or am I wrong?

[GenoMax]

If you have 48 cores then you would certainly want to up that num_threads option. You may need to experiment with that number to find the sweet spot for your hardware configuration. Start at the half-way point (and go up and down).

If you have admin access on the machine you may want to make a RAMDISK and copy the database files there (http://askubuntu.com/questions/15286...ake-a-ram-disk) If you are not using linux then look for an appropriate set of instructions for your OS.

[salamay]

Increasing the threads definitely helps. I guess I'll fiddle around with that. Thanks GenoMax!

[debem]

Hi ..
I am trying to get only the "no hits" result after perform the Blastx on the command line

I've tried:

<my Blastx command> | tee redirect.result33.txt | cut -d ' ' -f 1 | uniq | comm -31 - ids33_query.txt

Any help would be appreciated.

Thanks

[Rhewter]

Hi debem,

Input your results in to two files.csv and try:
sdiff 1.csv 2.csv > difer.csv

Open the file difer.csv and select the lines with "<".
This is your differents IDs.

Big hub,
Rhewter.