We compared control tumor and our gene of interest knock out tumor by RNAseq analysis and got very good comparative results. The problem is when we ran the qPCR using the cell lines derived from respective tumor types, expression is almost completely opposite and in some cases the minimal difference. I have ruled out mislabelling but this has completely stumped me. If anybody with better understanding can suggest a plausible explanation, I will be really grateful.
Thanks
Ravi
Hi Ravi,
Just to clarify, did you knockdown a specific gene of interest, in order to see what effects that had on the expression of many different RNAs? Then did you compare samples from the control tumour cell line vs. the gene-of-interest-knockdown tumour cell line using RNASeq first, and then secondly make the same comparison using qPCR in order to try to confirm what the RNASeq showed?
I have used the miRNA-Seq approach to profile changes in microRNA expression in cell lines following drug treatment (control cells vs. cells treated with the drug). My analysis of the miRNA-Seq data indicated that lots of microRNAs were either up- or down-regulated by the drug treatment. However, I then used RT-qPCR and/or Northern Blotting to try to confirm these changes. The RT-qPCR and Northern Blot data confirmed expression changes of some microRNAs, but there were actually many microRNAs that appeared to be modulated according to the miRNA-Seq data, but for which the RT-qPCR and/or Northern Blot experiments showed no change, even after I confirmed by dot-blot experiments that my probes were binding to the correct microRNAs. Thus it was clear that miRNA-Seq data did generate a number of 'false-positives'. Nevertheless, some interesting modulations were eventually able to be confirmed by RT-qPCR and/or Northern Blots. The generation of 'false-positives' in miRNA-Seq is inevitable due to the sheer complexity of the data generated, and the need for normalisation.
I would be curious to know how many RNAs appeared to be differentially expressed from your RNA-Seq data, and then how many of those RNAs you have tried to validate by qPCR so far.
Good luck!
Chris
Thanks Chris
We induce tumors with control mice and mice with knocked down gene, then we did RNAseq on both stem cell derived from tumors and tumor tissues which correlate very well. We looked at the pathways and genes involved in 2 different tumors and analysed the DE genes. We find the oncogenes which are differentially expressed and then did the qPCR to validate our RNAseq.
Ravi
Hi Ravi,
When I have run into similar situations, I usually go back and look at what the qPCR primers are actually targeting. From experience, RNAseq data (depending on how you analyse it) will give expression values for the WHOLE gene, and sometimes only part of the gene is lost, not the whole gene. However, since you mentioned oncogenes, I am going to assume you mean that your genes of interest are UPregulated in the DE analysis.
How are you performing your qPCR? Are you performing comparative CT or are you looking at fold changes? What are the controls you are normalising against? These are all things that can have an effect on results. Are you completely knocking the gene out? If so, I would expect no product on the qPCR. Are you performing a melt curve? I would definitely be looking at a melt curve very carefully because poorly designed primers may result in primer dimer or unexpected products, which will give a CT when you are expecting no CT.
As a first step I'd be performing the melt curve, and also running the product on a gel to confirm the single band.
Thanks
Sophie
Thanks Sophie for taking time out to answer, I understand what you are saying, I think I didn't made my question very clear. We knocked out the gene using Cre and it is verified and we seen definite phenotypic and morphological differences between 2 tumors. We are looking at other DE genes not the one knocked out to see what alternate pathways the KO tumor uses to sustain itself. We identified pathway which was upregulated in KO tumor compared to Control tumor. And I checked qPCR for this pathway components and targets. I used double delta Ct method with GADPH as control and also looked at the melt curve which are fine.
Ravi
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Cell Biology
- Culture Cells
- Cell Line