I am preparing samples for NGS, RNA sequencing.
I isolate total RNA from B. subtilis, deplete it from rRNA and then make DNA libraries. Every step is quality-checked using BioAnalyser.
And in the last step some issue occurs:
I have 12 samples and in case of 4 of them, the second peak after PCR amplification appears, suggesting large heteroduplex DNA complexes creation. So I lower the number of cycles in my PCR and everything works fine. But there is the question: Can we use samples with different numbers of cycles in amplification? Will be these samples comparable? Or is a better "scientific" way to have all samples with the same numbers of cycles, even though in some DNA libraries will be large heteroduplexes?
Thanks for your kind advice.
I would answer for two conditions:
1. NO, we should not do that and it is not preferred if you mix the samples with different number of PCR cycles. As you said, it is not a scientific way. The reason is, for any NGS, the PCR is a troublesome step as it induces bias i.e., different amplification for different templates. Thus, for every PCR run, its sort of unique amplification (practically). Moreover, it is highly recommended to manage all the samples in single/same PCR run. Therefore, now a days there are PCR machines with multiple heat blocks. Furthermore, developed NGS strategy would highly recommend to make the sequencing PCR free so that it would produce more accurate results for the relative abundance of different amplicons. So, in your case you cant compare the samples as there would be difference in the relative abundance due to PCR.
2. this is the sort of non-scientific way, yes you can mix the samples in NGS if you dont want to compare the relative abundance or quantification of the amplicons, and if you just wanted to see the presence of those amplicons. Hope this would help.
Theoretically after a certain level pcr product will reach a platue but because ngs is ever sensitive and we want to do good science Abhijeet is right you should do the same cycles for all your samples ...
And on a second note ... don't you have a contruct that are you supposed to sequence (i.e a certain band size that you can slice out and purify and sequence?)
If the answer to the second question is yes and the resulting amplified construct is of the correct size and good quality ... then why do you want to mess with the pcr cycles ?
Before even thinking about the pcr cycles i would first make sure the construct primer concentration is correct and exclude any pipeting error or contamination of my samples ...
Is there something much different about the 4 samples vs the other 8 that are without extra bands ( i.e quality , concentration, is rRNA depletion going equally well for all your samples etc ?)
Finally it would be much more preferable to optimize the number of cycles that would give no nonspecific bands for any of your samples and apply that to all your samples than do different cycles for some samples.
As long as you describe the number of cycles in your methods and your ngs quality controls and data results are good then you should be ok
Hope this helps and good luck
No because you would artifically inflate the abundance of your reads with more PCR cycles. Remember you are doing PCR to amplify reads so that you can get a detectable signal.
Thank you all for answers and suggestions!
I was expecting that changes in PCR cycles will be not-comparable.
Problem with my sample you can see in the attachment (profile of DNA library)
I discuss the issue with the GeneCore, Heidelberg and I finally get the answer - this second peak is not an obstacle for Ilumina based sequencing. Maybe it is only some PCR bias, which will not bind to the flow-cell chip or even though it is some DNA heteroduplexes, bridge amplification preferably works for shorter fragments. So I can use all my samples.
Sure, the Illumina sequencing is limited to only small fragments and it could only reads around 300bp and amplicon of around 600-650bp (2*300bp-pair end reads). Thus, major fraction of your reads would be the 300bp fragment. If there would any artefacts in sequencing, it would be already removed by the sequence processing step which will removed the reads without paired-ends. So, keep the PCR cycles same for all and it would be okey for your samples now.
I agree with GeneCore and previous answers. You should absolutely keep the same number of PCR cycles when you want to compare samples, mainly due to the important biases introduced by PCR, as said above. It's even a very good idea if they can all be sequenced on the same flow cell. As far as high weight DNA heteroduplexes, you may lose some read depth with those, but your libraries are not so much over-amplified. I've sequenced libraries where both peaks were at the same height and still got some reasonably good results (beware, I do NOT recommend to over-amplify NGS libraries, I'm writing that it's not so problematic if you do). So I would definitively go with your first round of libraries and sequence them. Good luck!
- Badges
- Science method