Merging blast databases

Question

[Joe]

Is it possible to merge 2 blast databases?

Say I want to query NR, but at the same time would like to query some sequences of my own that are not yet in nr (or lets say I'd published them by the local copy of nr was out of date and I don't want to download it all again).

I don't suppose it would be as simple as just concatenating the 2.

[Pierre Lindenbaum]

how about running two blast queries for each database ?

[Michael]

That is not that good of an option, because that will not adjust the database size to the compound db size, therefore e-values will be incorrect and are incomparable, when you are in fact doing "single query in two passes". Making an alias will adjust to the new database size. Also, combining the blast output in some formats is not trivial.

[Joe]

That is currently what I'd be doing. I'm kinda after having all the results together in the same file. I could do the after the fact and merge the result files, but if it was possible to merge DBs that seemed preferable to begin with

Answer 1

[Michael]

blastdb_aliastool is the command designed for this purpose. See its help pages. It is even easier than concatenating two files because there will be no third large compound file, but only the alias file.

[Joe]

Ok great, I'll take a look - cheers!