small RNA

[snorider]

We are getting ready to run small RNA samples. Illumina recommends
running a control lane for these types of samples. Is it possible to
spike in PhiX to each lane instead of using a whole lane for the
control? If it is what are the analysis issues? How are the PhiX reads
separated from the small RNA reads? Is there a chance that PhiX reads
can map as small RNA or vise-versa?
Thanks,
Bruce

[Michael Muratet]

Bruce

I believe what Illumina is recommending affects upstream processing.
Including a (PhiX) control lane allows the pipeline image processing
parameters to be calculated over a sample that has balanced nucleotide
content. I don't think it's about calibrating expression (but someone
please correct me if I'm wrong) so you don't want to mix it with the
sample(s). If you don't set a control lane the image parameters will
be calculated over all eight lanes and if the samples are biased WRT
content you can get skewed results. I don't know that anyone has every
taken a hard look at why this is so, but several groups have proposed
alternatives to the Illumina processing that offer improved
performance starting from different places in the process. There have
also been a number of papers looking at biases in next-gen sequencing.

The pipeline has always allowed a control lane. The last version
(v2.4) of SCS did not. I forget what the IPAR allowed. We don't use
phiX control lanes in production here but we do include a lane we
think has balanced content, e.g., a RNA-seq lane, for a control.

Depending on how many flow cells you plan to do, you could do one with
the phiX control and then repeat the processing without the control
lane or with a sample lane as the control to see what difference it
makes.

Good luck

Mike

> Thanks,
> Bruce
> >

Michael Muratet, Ph.D.
Senior Scientist
HudsonAlpha Institute for Biotechnology
mmur...@hudsonalpha.org
(256) 327-0473 (p)
(256) 327-0966 (f)

Room 4005
601 Genome Way
Huntsville, Alabama 35806

[elaneyk]

Hi Michael,

I'd be wary of using an RNA-seq sample lane as an example of a
balanced content lane. We've always seen a % bases called skew up to
cycle 15 in RNA-seq samples from human and non-human RNA sources and
personally I'd never use it in place of a balanced genomic DNA sample
like PhiX (have a look at the s_N_percent_all.png plots in the Bustard/
Plots folder and compare RNA-seq vs PhiX/genomic DNA) . We recall all
our bases for RNA-seq data using a PhiX lane as a control and ignore
the bases called by RTA .

Elaine

> mmura...@hudsonalpha.org

[mitch]

Hi Bruce,

I would recommend using a control lane by itself to be on the safe
side as you do need to allow for the uneven base composition of the
small rnas. Spiking in the phiX would also impast on the alignment if
you were relying on eland for the alignment stats. The new scs2.5 and
rta software now allows you to nominate a control lane so the base
calling can be done as soon as the run has completed the image
processing. Without this new software you would have to re-call the
bases using pipeline.

Cheers,

Mitch.