Question: Will large differences in the number of reads between samples decrease reliability of DESeq2 analysis?
gravatar for Mike
3 months ago by
Mike10 wrote:

I have some RNA-seq samples from mouse, 2 conditions with 4 replicates each, read quality is good and for each sample 85-90% of reads align. The number of aligned reads in millions are:

Condition 1 replicate 1: 100
Condition 1 replicate 2: 79
Condition 1 replicate 3: 52
Condition 1 replicate 4: 37
Condition 2 replicate 1: 56
Condition 2 replicate 2: 59
Condition 2 replicate 3: 31
Condition 2 replicate 4: 26

I know DESeq2 normalizes based on the number of reads but I'm wondering if the variability here will cause any problems?

ADD COMMENTlink modified 3 months ago by Ryan C. Thompson6.3k • written 3 months ago by Mike10
gravatar for Ryan C. Thompson
3 months ago by
The Scripps Research Institute, La Jolla, CA
Ryan C. Thompson6.3k wrote:

This looks like a reasonable, expected variation in total read counts. All the totals are within one order of magnitude. I don't think the differences in sequencing depth will be a problem for this data. I've used similar methods successfully on data sets with much wider ranges of total counts.

ADD COMMENTlink written 3 months ago by Ryan C. Thompson6.3k

Thanks that's good to know.

>This looks like a reasonable, expected variation in total read counts.

We sequenced on a NextSeq and expected the reads evenly distributed between the 8 samples and I was surprised by the variability, I expected some but not this much. I was planning to follow up with the sequencing facility because I was a bit concerned that the variability could be due to errors in quantification or poor quality control by them (we submit RNA and they do the rest). But you think the numbers seen here are normal?

ADD REPLYlink written 3 months ago by Mike10
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 168 users visited in the last hour