bowtie2 very_sensitive command in RSEM-1.2.11
j.mi...@ec.gc.ca
- --bowtie2
-
Use Bowtie 2 instead of Bowtie to align reads. Since currently RSEM does not handle indel, local and discordant alignments, the Bowtie2 parameters are set in a way to avoid those alignments. In particular, we use options '--very-sensitive --dpad 0 --gbar 99999999 --mp 1,1 --np 1 --score-min L,0,-rate'. "-rate", the last parameter of '--score-min' is the negative value of the mismatch rate provided by option '--bowtie2-mismatch-rate'. If reads are paired-end, we additionally use options '--no-mixed' and '--no-discordant'.
However, I recently tried using using the bowtie2 option in RSEM 1.2.11 and noticed that the preset being called is "sensitive" and not "very-sensitive" as stated above. And looking at line 68 of the rsem-calculate-expression perl script:
my $bowtie2_sensitivity_level = "sensitive"; # must be one of "very_fast", "fast", "sensitive", "very_sensitive"
Just thought I would point that out. I'm trying to figure out why my (bowtie/bowtie2) overall alignment rate dropped from 81.12% with bowtie to 49.03% with bowtie2. I'm aligning 3 million 454 cDNA reads (single end) to a de novo transcriptome. I used the RSEM defaults for both bowtie and bowtie2, and specified -fragment-length mean and sd.
j.mi...@ec.gc.ca
- --bowtie2-sensitivity-level <string>
-
(Bowtie 2 parameter) Set Bowtie 2's preset options in --end-to-end mode. This option controls how hard Bowtie 2 tries to find alignments. <string> must be one of "very_fast", "fast", "sensitive" and "very_sensitive". The four candidates correspond to Bowtie 2's "--very-fast", "--fast", "--sensitive" and "--very-sensitive" options. (Default: "sensitive" - use Bowtie 2's default)
I also missed the fact that the sensitivity can be changed using the -bowtie2-sensitivity-level setting. guess its called a README for a good reason.....
Sorry about that
Bo Li
Thanks a lot for pointing this out for us. This is a typo on the
'rsem-calculate-expression' document. I'll fix it for the next release.
Best,
Bo
On 2014-03-04 11:54, j.mi...@ec.gc.ca wrote:
> In the help file for rsem-calculate-expression it states:
> RSEM website: http://deweylab.biostat.wisc.edu/rsem/ [1]
> ---
> You received this message because you are subscribed to the Google
> Groups "RSEM Users" group.
> To unsubscribe from this group and stop receiving emails from it,
> send an email to rsem-users+...@googlegroups.com.
> To post to this group, send email to rsem-...@googlegroups.com.
> Visit this group at http://groups.google.com/group/rsem-users [2].
>
>
> Links:
> ------
> [1] http://deweylab.biostat.wisc.edu/rsem/
> [2] http://groups.google.com/group/rsem-users
Bo Li
Thanks for sharing these interesting results with us. I guess that the
allowing of gaps might be a good reason for the difference. But I'm not
sure if the setting of --score-min will have such a big difference.
Please note that although the default setting of --score-min is
L,-0.6,-0.6 but RSEM uses L,0,-0.1, RSEM also sets --mp to --mp 1,1
(default of bowtie2 is --mp 6,2).
Best,
Bo
On 2014-03-05 06:54, j.mi...@ec.gc.ca wrote:
> BTW I changed the setting in the rsem-calculate-expression perl script
> to "very-sensitive" and still had low alignment overall (49.05% of
> reads vs 49.03% using the "sensitive" setting)
>
> When I use Bowtie2-2.2.0 directly (on a similar set of 454 fastq
> files) , --very-sensitive and otherwise default settings, I get around
> 82% alignment. Mind you this is allowing for gapped alignment
> (probably a good thing with 454 reads), but bowtie (ungapped)
> alignment on the same run files was not that much lower, around 75%. I
> wonder if maybe this difference is due to the --score-min function,
> which by default is L,-0.6,-0.6 but in RSEM is L,0,0.1. meaning that
> for a 500bp 454 read the min acceptable alignment score by default is
> -300.6, and in RSEM is 50 (if I understand correctly).