Dear Shujun,

When comes the time to deal with very low amount of RNA, I would suggest you to be very careful with contamination that is unavoidable. If you are using regular Eppendorf tubes, some DNA/RNA/protein will stick to it on the borders. I normally use Eppendorf LoBind tubes to reduce the bias due to what sticks in the tube. Normally I get better results that way.

Hope that was useful!

You should Continue to work from A to Z without storing in deep freezer precisely , with optimization in concentration

best wishes 

For smaller amount of cells kits perform better vis-a-vis Tri reagent. I sthere any chance you could try one like RNAqueous, Qiagen RNeasy 

I think what Shujun is trying to say is that she gets RNA signal in samples to which she added no RNA in the first place.

Shujun, my guess is that you're using nanodrop for RNA quantification. As good as glycogen is in facilitating the precipitation of low quantiities of RNA, unfortunately, glycogen itself absorbs UV at wavelengths similar to nucleic acid. Therefore, in your no RNA control you will get measurable absorbance in the 260-280nm. On top of that, the glycogen absorbance curve is different to that of nucleic acids, it will skew the absorbance from the RNA itself and it will appear in your read-out as contamination. 1000 cells is indeed a tiny amount and I don't think you can get enough RNA out of it to be able to measure it reliably with nanodrop. What you can use is a more sensitive measurement lik Qubit. Otherwise, you can use qRT-PCR that includes a standard curve, but that uses up your sample and requires reverse trancription step.

You can also try using the glycogen-only sample as the blank for your measurements. This way, whatever absorbance is caused by glycogen, it will be subtracted from the RNA's absorbance. But at very very low RNA concentrations, you might still only get some noise rather than a nice curve.

Good luck!

Thank you guys so much for your kind suggestions. I am still trying!