BLAST IN CLC vs BLAST IN NCBI website

Question

[lukhanyomakhabane]

Hey guys. I'm planning to do BLAST searches for some gene sequences of interest in several fungal genomes I'm working on. some of these genomes are available on NCBI database but others are not available, hence I retrieved them from other sources. So now I was thinking of doing BLAST searches on NCBI websites for those genomes available there. Hereafter, I do Blast searches on CLC for those genomes not available. So my question is; would this cause some biase in my results if do BLAST in separate platforms NCBI websites vs CLC genomic Workbench. Or maybe I should just conduct the BLAST in CLC, like download the genomes on NCBI and use them in CLC. I'm not sure if my question clear, your input and suggestions would be greatly appreciated.

Best wishes

Answer 1

[lieven.sterck]

Yes, that will create a bias, question is whether that will be a problem.

Blast statistics are calculated using, among other numbers, that database size, so if that changes your obtained values (mainly the E-value) will be biased. On the other hand that will only be a problem if you want to compare results from those blasts using such statistics, so directly compare if an e-value of blast A is better/Worse than e-value from blast B . The best/top hit will still be the best hit regardless of which blast you ran but the statistics will differ.

Two options:

• you indeed run all those blasts in the same way, same software against the same (local) DB
• with NCBI blast you can manually set the DB-size. Using that you can tell blast to calculated stats making use of the give size rather than the effective DB size you are using.

side note: I'm also not sure NCBI blast and CLC blast use the same algorithm so more differences or discrepancies can occur due to that.

[lukhanyomakhabane]

Thank you for this information!