Velvet "Strand_Specific" Argument

Question

[Newvin]

The Velvet manual documents a "strand_specific" argument for transcriptome assembly. If this option is specified, how does the assembler ascertain the strand that a read comes from? My guess would be it looks for a trailing "/1" or "/2". Can anyone confirm this or give me a better answer. Here is what the Velvet Manual has to say on the issue:

"Strand specific sequencing

If you are using a strand specific transcriptome sequencing protocol, you may

wish to use this option for better results. Even though it can be used anywhere

on the command line, it applies to all the dataset. To toggle it on, type:

./velveth output_directory/ 21 (...data files...) -strand_specific"

Thanks!

Answer 1

[SES]

There is a similar option with Trinity for assembling strand specific libraries. This option tells velveth that your data is strand specific, so your input data should be F (/1) or R (/2) but it should not be both. If you have paired-end data that is not strand specific, then it would probably be best to just use the normal -shortPaired option and leave off the "strand_specific" flag. I say that because I can't think of any advantages to assembling with one pair if your goal is to assemble the whole transcriptome. If you do have strand specific data, then I would assemble the sets separately with the "strand_specific" flag, that way you can compare strand specific differences. Hope that helps, maybe someone else can chime in, as well.