Question: two transcriptomes, same RNAseq, different results. why?
gravatar for rm5585
2.9 years ago by
rm55850 wrote:

Hi all,

I am trying to figure out the best way to make a list of genes that are differentially expressed after gene knockdown. RNAseq was done on a set of samples, about 30million reads/sample. 270 million total reads. I mapped the reads to transcriptome #1 and made a list of candidate genes FC <2, p-val<0.05. This resulted in a list of 286 genes. I did the same thing for transcriptome #2 resulting in a list of 315 genes. Both transcriptomes are from two different labs but are on the same wildtype animal. I expected that the two resulting lists would have been very similar. However, when I compared those two lists and only about 1/3 of the genes were the same between the two. Is there any reason why this might be?

Thank you

transcriptome rna-seq • 819 views
ADD COMMENTlink written 2.9 years ago by rm55850

Do you replicates for the individual samples? You could pool them all together then redo your analysis that way.

And yes this (getting different results on the same type of sample) is not unexpected and only underscores that biological variability, sequencing and sample preparation can have strong effects on the results.

The common genes are more likely to be correct, yet at the even the transcripts found in only one experiment might still be valid.

ADD REPLYlink modified 2.9 years ago • written 2.9 years ago by Istvan Albert ♦♦ 83k

Yes, I had replicates for all samples. What do you mean pool them all together?

Thanks for your help.

ADD REPLYlink written 2.9 years ago by rm55850

And how many biological replicates per sample per experiment? With few replicates, you will have low power to detect DEGs, and low overlap between two low-powered studies is to be expected.

See Schurch et. al 2016 for a good exploration of sample size and power in RNAseq.

edit: if the two experiments are independent, but using the same samples, you could analyse them together, adding a factor to control for batch effect.

ADD REPLYlink modified 2.9 years ago • written 2.9 years ago by h.mon29k

There are some details that are not clear to me: 1) FC <2, p-val<0.05, I guess it is: absolute value of FC > 2, p-val < 0.05 2) What do you mean by transcriptome #1 and #2? Did you get different transcriptome assembly? OR are two different RNA-seq experiments? 3) The experiments are on the same wildtype animal, but where do the knockdown RNAseq come from? Did you get both the before and after conditions from two different labs? Same animal, same treatment, same RNA extraction? Apart for this doubts, I very much agree with Istvan. However, if the experiments were on the same RNA extraction, I would start being a little bit worried, because the difference is due to technical variability

ADD REPLYlink written 2.9 years ago by Fabio Marroni2.5k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1020 users visited in the last hour