STAR.index for small genomes

[yeroslaviz]

in order for STAR to work correctly with smaller genomes, it is recommended to change the --genomeSAindexNbases parameter while indexing.

Is this possible from within the package and this command, or do I need to give it a script?

If needed I can also modify the internal Orfik script, but I don't think it is a good idea.

thanks

[Roleren]

Ignore this warning unless it crashes, the modern version of star usually handles this very well. Did it actually crash or just give a warning?

[yeroslaviz]

No, the indexing went well, and this was a warning only.

but it crashes during the mapping with the following error

EXITING because of FATAL input ERROR: --limitIObufferSize requires 2 numbers since 2.7.9a.
SOLUTION: specify 2 numbers in --limitIObufferSize : size of input and output buffers in bytes.

I'm using STAR 2.7.10a

In order to overcome this, I have modified the lines in the script RNA_Align_pipeline.sh script to be as such:

--limitIObufferSize 30000000 50000000

I'm testing it now.

[yeroslaviz]

Just wanted to say, that changing the --limitIObufferSize parameter fixed this issue.

[Roleren]

Yes, so here is a small problem in a sense, because ORFik supports STAR 2.7.4a. So this is more a request for enhancement, will see if I have the time to add it in. If not for you, just use the RNA_Align_pipeline.sh each time. Your hassle now is that you need to remember to use it each time, while with 2.7.4a it just works :)

[yeroslaviz]

Not sure if it will continue working, but for simplicity, I have put the modified file in the same STAR_Aligner folder and it seems to work fine for now. let's see how it goes.

I'll close it for now, if it is ok.