32 is ok, but 36 is not ok.

There can not be any huge difference if you use 600 ng or 1 µg. This should actually make less than 1 cycle difference.

Try NOT to dilute the cDNA. This should give you a shift of about 3 cycles (what will shift 36 to about 32, what would be useable).

Having lower Cq values for the GOI is not a problem (given the general requirement that the amplification efficiencies for the GOI and the reference genes are identical). It's a sign that there is a hell lot of viral RNA. This is possible.

Chosing good reference genes for viral infection studies is difficult because many viruses induce a "transcriptional shut-down", at least to some extend, and not all genes are affected similarily. However, you could detect this by comparing the Cq values of the reference genes between infected and uninfected aphids.

If you want a reference gene with higher expression (lower Cq values) you might consider using rRNA as reference.

32 is ok, but 36 is not ok.

There can not be any huge difference if you use 600 ng or 1 µg. This should actually make less than 1 cycle difference.

Try NOT to dilute the cDNA. This should give you a shift of about 3 cycles (what will shift 36 to about 32, what would be useable).

Having lower Cq values for the GOI is not a problem (given the general requirement that the amplification efficiencies for the GOI and the reference genes are identical). It's a sign that there is a hell lot of viral RNA. This is possible.

Chosing good reference genes for viral infection studies is difficult because many viruses induce a "transcriptional shut-down", at least to some extend, and not all genes are affected similarily. However, you could detect this by comparing the Cq values of the reference genes between infected and uninfected aphids.

If you want a reference gene with higher expression (lower Cq values) you might consider using rRNA as reference.

Are you sure that your reference gene is not influence from the treatment or the study?

Jochen Wilhelm,

Thank you for tha fast reply!

The amplification efficiencies do differ. I'm trying to have them between 90% and 110%. For example: NADH and B-actin have efficiencies around 101% while my GOI has an efficiency of 96 % (and even 86% if I use another primerpair to detect the GOI).

Thanks for the information! I didn't think about the "transcriptional shut-down". For the moment I found it difficult to find uninfected aphids. I did found one aphid which had a very low viral amount but I didn't notice a big difference in Cq-values of the reference genes (Max 1Cq-unit).

So I don't know if completely non-infected will give very different results than?

I will surely try to find other reference genes, including rRNA

Giorgia Pertile

Well before yesterday I would have answered: Yes, because I don't do any treatment on my aphids. I just take them from the field en immidiately freeze them and check if they are infected with the virus or not. But maybe the viral infection does influence the reference genes?

Is possible that this infection influence your reference gene. I agree with the other comment that is best to try to test more gene and after decide which is the best.

Standard questions:

What, exactly, is NADH (the gene) here? NADH dehydrogenase? If so, I think we can largely rule out transcriptional downregulation, since both NADH dehydrogenase and beta actin are high-abundance (and essential) transcripts. A downregulation of this magnitude would likely result in a dead aphid, very quickly.

How quickly were your samples frozen, and in what? And how were they stored? In my experience, beta actin in particular is a high-turnover transcript (decays rapidly after death). Snap freezing in liquid nitrogen is better than slow freezing, and storage at -80 is better than storage at -20.

How clean is your RNA? You spec it, presumably (since you list a concentration) but do you check the 260/230 ratio? This is very important, particularly with guanidium-based RNA isolation methods (which Trizol is), as high absorbance at 230 indicates carryover of guanidium salts. These are chaotropic and generally bad news for delicate enzymes like reverse transcriptase, so too much guanidium contamination will massively reduce the efficiency of your cDNA synthesis, and not necessarily in a uniform fashion.

Given you are getting low yields, you are presumably using fairly large volumes of your RNA preps in each cDNA synthesis, so even modest salt contamination can be detrimental (this is why it's usually expressed as a ratio rather than an absolute value).

260/230 ratios of 1.7 or more are usually ok, but 2.0+ is ideal.

Is the DNAse step really necessary? It's an additional source of variation, it introduces an enzyme which destroys DNA (which could be bad for cDNA synth/PCR if you don't absolutely get rid of all of it), and it introduces buffer components which may or may not be detrimental to cDNA synthesis. It's especially risky when you have so little RNA.

Genomic DNA contamination is absolutely an issue, but you can design primers that span introns to avoid this issue, and if that's not possible, you can also check for gDNA contamination: do a few "no reverse transcriptase" controls, and if your Cq values for 'no RT' samples are ~6 or more cycles higher than your +RT samples, then gDNA contamination is present but negligible (pipetting error contributes greater variation than gDNA, under such scenarios).

Random hexamers vs oligodT: have you tried oligodT? I realise some viral RNAs may not have polyA tails, but you can use both in combination to maximise your cDNA synthesis -as long as you do this consistently, it's fine.

Sample quantity: have you tried just prepping RNA from a big bunch of aphids? Collected maybe a hundred and mash them all up in trizol together to generate an RNA pool that doesn't have all the issues that 'low RNA content' brings: then validate your reference genes using this stock, which you will

A) have lots of, and also

B) won't particularly mind using up, because it exists only for testing purposes

You could also check a number of other genes in this stock, to see if your abundance issue is global or gene-specific. And yes, checking for something like 18S is a very good idea. Ribosomes are ~85% of a total RNA pool, so the Cq values for 18S should be very low.

Finally, before I forget, well done for using multiple reference genes (many investigators do not), good job on diluting your cDNA (cDNA synthesis buffer can disagree with PCR polymerases), and absolutely top marks for recognizing that very high Cq values may be indicative of a problem that needs addressing. Big thumbs up.