September 13, 2009


Notes on JGAST use:

(1) JGAST is somewhere in between a local alignment algorithm (such as BLAST) and a global alignment algorithm (such as CLUSTAL). If a query sequence and a target sequence are on opposite strands, BLAST will reverse one of the sequences so that both sequences are on the same strand. CLUSTAL will not make this correction. In it’s current iteration, JGAST does not make this correction. That is, JGAST expects your query sequences to be in the same 5’-3’ orientation as your target sequences. This will usually be true for 454 pyrosequencing (in which all the sequences will be in the same 5’-3’ orientation), but is not generally true for Sanger sequencing of a clone library. I have code that will take 16S sequences and put them all in the same 5’-3’ orientation. If you would like that code, let me know (at anthony.fodor@gmail.com). If there is enough demand for it, I will incorporate it into JGAST.

In summary, please ensure that if you use JGAST your query sequences and target sequences are in the same 5’-3’ orientation.

(2) JGAST starts with a blast search and the blast results are stored using BLAST’s "-m 8" option. This is a tab delimited file. If your FASTA file headers (the lines that start with ">") have tabs, and those tabs get copied over to the BLAST output, this may confuse the Java program that parses the BLAST output. Please ensure that there are no tabs in the headers of your FASTA file. I can send you code to purge your FASTA files of tabs. And if there is enough demand, I will add this code to JGAST.

If you have any questions, please contact me at anthony.fodor@gmail.com