Hi Boustani,

RNA extraction from plant material can often be contaminated with carbohydrates, which can be one of the reasons for low 260/230 ratio and unfortunately, simple ethanol ppt will not solve the issue. Also, your 260/280 does not look very optimistic too.

What extraction method are you using? If you are using a kit with column, overloading the column with tissue can results in low yield and bad ratios (trust me, being greedy is not good). If you have the columns, you might run the RNA through them and clean it up per manufacturer's instruction.

Another way is to go through PCI again, which ironically should help to remove the organic contamination with good cleaning. Of course, the choice of cleaning method would depend on how much RNA you have to play with. If you have abundance of tissue, I would suggest you start over but identify steps in your extraction method that could help to reduce all types of contamination.

Since you are using the RNA for qPCR, you want to make sure the quality is reasonable good before you proceed. So, if you have access to Bioanalyser, you might want to run it through and see what you get. The RIN seems to be a good indication for whether you will get good qPCR results. I am attaching an article that published this (apart from the fact that this is an article from Agilent itself.....) If not, do run a denaturing gel to make sure your RNA is still intact.

All the best!

http://www.biomedcentral.com/content/pdf/1471-2199-7-3.pdf

How low is low? And how does the 260/280 look like?

Phenol residues will interfere with the reverse transcriptase.

I have two groupes :

1/ the lowest ones 260/230 : 0.3 to 0.9 (Leaves and Petioles RNA)

they have 260/280 : 1.10 to 1.60

2/ the highest ones 260/230 : 1.26 to 1.72 (Root and fruit ARN)

260/280 : 1.9 to 2.02

If you are comparing the expression of a gene among all these organs (root, leaf, fruit, petiole), you may want to have similar RNA extraction efficiency and pretty much same quality; however, after you synthesize cDNA (if phenol contamination is not a problem for cDNA synthesis), if your endogenous control gene is consistant in all organs, I guess your gene expression analysis should be valid :)

I did organ specific expression analysis of five genes, I know it is really difficult to extract same quality RNA from all organs (root, stem, leaf, flower, silique) in one batch; I remember I had to extract RNA 4/5 times to get there :) All the best for you!

In my experience, the ratio is dependent of RNA concentrations and is usually low in concentrations lower that 100 ng/ul. I would worry only if you have higher concentrations.

Hi Boustani,

RNA extraction from plant material can often be contaminated with carbohydrates, which can be one of the reasons for low 260/230 ratio and unfortunately, simple ethanol ppt will not solve the issue. Also, your 260/280 does not look very optimistic too.

What extraction method are you using? If you are using a kit with column, overloading the column with tissue can results in low yield and bad ratios (trust me, being greedy is not good). If you have the columns, you might run the RNA through them and clean it up per manufacturer's instruction.

Another way is to go through PCI again, which ironically should help to remove the organic contamination with good cleaning. Of course, the choice of cleaning method would depend on how much RNA you have to play with. If you have abundance of tissue, I would suggest you start over but identify steps in your extraction method that could help to reduce all types of contamination.

Since you are using the RNA for qPCR, you want to make sure the quality is reasonable good before you proceed. So, if you have access to Bioanalyser, you might want to run it through and see what you get. The RIN seems to be a good indication for whether you will get good qPCR results. I am attaching an article that published this (apart from the fact that this is an article from Agilent itself.....) If not, do run a denaturing gel to make sure your RNA is still intact.

All the best!

http://www.biomedcentral.com/content/pdf/1471-2199-7-3.pdf

You can determine whether low purity is a problem by running qPCR on serial dilutions (1:10, 1:100, 1:1000) of your sample. At good efficiency , CT values for each dilution should be about 3 cycles apart. If it is less than this, inhibitors may be present in your sample. More details are available in the link below.

I second the advice above for using a Bioanalyzer (or equivalent) to check the integrity of your RNA.

From what species are you isolating RNA? I have some experience with this and may be able to recommend a method that will improve your purity. As mentioned above, too much tissue is bad for any kit/protocol.

http://www.invitrogen.com/site/us/en/home/Products-and-Services/Applications/PCR/real-time-pcr/qpcr-education/real-time-pcr-troubleshooting-tool/gene-expression-quantitation-troubleshooting/poor-pcr-efficiency.html

Just a detail: too often the statement about the Cq difference expected from serial 1:10 dilutions of samples used in qPCR/RT-qPCR is said to be "about 3" or "about 3.3." A direct way to express this value is 1/(log(base10) of 2) which = 3.321928... So, if it ever becomes popularly and/or permanently forgotten what this value [technically] is, and where it came from, all one has to do is remember that it is simply the "reciprocol of the log(base10) of 2." Or, perhaps even more succinctly: it is the "log(base2) of 10"

...

- Which opens up the entire relationship between efficiency and expected Cq differences between samples along a known dilution series (and, ultimately, this is also the math that is being applied to the differences in Cq values between and among experimental unknowns/samples themselves):

"log(base Eamp) of the dilution factor = the expected difference between adjacent Cq values along that dilution series"...

...

*E.g. (Assuming that no reaction inhibition is present) for serial 1:4 dilutions at 80% amplification efficiency, this equation becomes: log(base1.8) of 4 = Cq differences expected = 2.358 ... and so on for every efficiency (here called **Eamp for "Exponential amplification") and every/any possible dilution factor.

...

This also underscores the large effect that Eamp has on quantifying gene expression. AKA: It is highly important to assess amplification efficiency - but over many same-target reactions, not singularly. Then averaging those Eamp values, not arithmetically, but in correct, log-based fashion: E.g. if Eamp values of 2, 1.95 and 1.87 were observed for a target (either by multiple standard curves or by multiple valid [uninhibited] same-target reactions), the average is not a simple arithmetic average (which would arrive erroneously at exactly 1.94 here) ... ... rather it would be the following equation [in log(base10) vernacular]:

10^(1/Average(1/log(Eamp1),1/log(Eamp2),1/log(Eamp3))), which would = 1.93698... (a small difference in comparison to "1.94," but, nonetheless, more theoretically correct).

*The above example of using serial 1:4 dilutions is one way to asses target amplification efficiency and standard curve behavior.

**Eamp = (Efficiency + 1), so, at 0.8 (or 80%) Efficiency, Eamp = 1 + 0.8 = 1.8

Would be much better if people would use 1:2 dilutions for this purpose.

To inspect more slices of (and examine the exact contours, behavior and limits of) the 'entire' valid dynamic range for each target standard curve - I agree entirely.

Yes, and I additionally meant that then it is clear and simple that there should be 1 cycle difference between adjacent dilutions (given optimum efficiency).

And it is very near (or at) the limit of believable resolution of the assay as well...

When no technical replicates are used, yes.

Which brings up the importance of running the right number of technical replicates during use of nil-reaction (Poisson-modeled) analyses to determine if one has identified a single-copy reaction or not. E.g. number of copies = -Ln(#nil rxns/#total rxns), wherein ~6 (6.29) positive rxns and ~10 (9.57) nil (no signal) rxns out of 16 total reactions would indicate a single copy of target. In such studies, 48 total rxns [tech. reps] would be best; but 16 is thought to be the useful/informative minimum # of total rxns to use. Nil-reaction analyses provides a way to make some sense out of the 'schizoid' nature of Cq values in the dilute (single-copy) region. It's good to have an idea of what Cq spread one can expect from a single copy rxn for every target as it opens the door for an approach to near-absolute copy # quantification. So, agreed, a very high extent of resolution can be achieved by appropriate use of technical replicates as long as the lab's bank account holds out ;-] Many lab's are running on fumes these days (including here) - so, unfortunately, some of the corners that are cut sometimes include qPCR/RT-qPCR technical replicate numbers - even when running 15 uL-size reactions. In a good corporate setting, the sky should always be the limit. But, in soft-money-research academia, the current wolves of financial distress are howling loudly out on the moors.

I extract rna from drosophila embryo wich are very hard to obtain i usually obtain 2ug in 20ul my 260/280 ratio is around 2 but my 260/230 is between 0.4 and 1. I never have problem in downstream application my biological replicate are reproducible. I can't use rna extraction kit because my rna yield is too low.

So as it is writen in qiagen website phenol contamination is not a problem for qpcr reproducibility.

I hope it is the case foe zverybody.

Jeremy