RNA/Sample storage experiment II - results

It turns out that it doesn't really matter when I freeze the sample, I can get high-quality RNA either way:

All 6 of my samples came out with really high RIN numbers. Talking with J~, it seems that the best way to go is to freeze the sample after the phenol-chloroform separation stage. At this point the RNA is pretty clean, certainly more so than in the RNAsolv buffer and I haven't used any of the columns or anything. Further, we're unclear of how the phenol-based RNAsolv buffer holds up for long-term freezing (this was only frozen for 2 days), whereas the top separated layer should remain stable frozen indefinitely.

Now, on to my samples!

RNA/Sample storage experiment

Before I launch into extracting the remaining 28 placentas, J~ wants me to run through a quick experiment:

The RNAextraction kit that I'm using (this probably holds for most RNA kits actually) calls for a small amount of tissue (15-50mg). As I am extracting placenta which routinely weigh up to 500mg (0.5g) I have a ton of extra tissue. I have an issue though, as placentas are highly complex: composed of at least 3 different cell types arranged in layers. This means that I cannot just cut off a small bit of placenta and use that (it isn't a random sample of the entire tissue - maybe I would end up with just 1 layer etc...). Instead I have to homogenize the entire sample first (normally with liquid Nitrogen) and then use part of the homogenized tissue. This sucks because if I mess something up, I ruin the entire sample.

Aside from the risk of ruining my samples, it sounds pretty straightforward. But it's not. Once a tissue is homogenized it has a much higher risk of RNA degradation (all the RNAses have been released from their cellular confines) and cannot just be put back in the freezer. Instead I add it to the first buffer of the RNA kit - in this case it's called "RNAsolv." The volume of buffer depends on the weight of the tissue  (in this kit it's 1ml/30mg) so with a 500mg placenta, I will end up with 500mg*1ml/30mg=~17ml of placenta-buffer slurry and I only need 1ml of buffer (30mg of tissue). Therefore I am left with obscene quantities of blueish-brown placenta slurry.

The question is: can we do anything with this slurry? And more specifically: is there any way to keep some of this around so that if I do mess up an RNA extraction, I can go back to the slurry mixture and just re-extract RNA instead of loosing the sample entirely?

To test this I will work through the extraction and at periodic points, freeze some of the sample, then in a couple days, thaw it out and finish the extraction. Here's an outline of the experiment:

This way I have 3 treatments: (1) "never frozen", (2) "top layer frozen" (freeze the liquid layer of the Phenol-Chloroform extraction), and (3) "buffer frozen" (freeze the RNAsolv - phenol - buffer right away). Each Treatment will be duplicated to get a sense of the variance.

I have just finished the extraction from #1 and I will finish up #2 and #3 on Wednesday. Then they can go on the bioanalyzer and we'll see if it works.

And, just because concentration is so important, here are the nano-drop values from these samples:
Never froze 1:       487ng/ul
Never froze 2:       546ng/ul
Top layer froze 1:  564ng/ul
Top layer froze 2:  530ng/ul
Buffer froze 1:      570ng/ul
Buffer froze 2:      549ng/ul

RNA extractions and the Bioanalyzer III: the solution

I finally figured it out. Naturally it was one of those mistakes that are considered "dumb" in hindsight, but it sure caused a ton of stress at the time. The problem was that I was not diluting my samples enough when I put them on the bioanalyzer. It's nothing to do with ethanol or chloroform or any of the other variables I spent the last couple months testing. I have been doing RNA extractions just fine the entire time, and then fucking up the quality assay.

Below you can see the differences. On the left there are two samples that I ran on October 1st. The desired concentration is between 200 and 5,000pg/ul and these were at a concentration of 92,448pg/ul and 54,308pg/ul. I diluted them down to the proper concentration and re-ran them yesterday (right column) and they turned out really great (RIN>8 is acceptably good, >9 is perfect).

So, what have I learned? The bioanalyzer is incredibly accurate in it's small range of acceptable concentrations, once you get outside of that range the quality goes way down. The nanodrop is mildly accurate, but can handle a huge range of concentrations. The first step is always to use the nanodrop to get a good idea of the starting concentration of a sample, then dilute the samples to fall within the range of high accuracy of the bioanalyzer. 

There are actually two types of chips for analyzing RNA with the Bioanalyzer: "Nano chips" and "Pico chips." The differences is the concentration they deal well with. Here's a picture:

                                           

Green = Nanodrop

Red = Bioanalyzer Pico chip (20-5000pg/ul)

Blue = Bioanalyzer Nano chip (5-500ng/ul)

It turns out that if I had been using the Nano chips, my RNA probably would have fallen perfectly within the range. That would have saved a bunch of time, heartache, and money, but on the other hand I wouldn't have learned the concentration lesson in such a lasting and exquisitely painful manner.

I now have nano chips which I will be using for my samples and I am going to run the rest of my samples to see whether I ruined them (as I had thought before) or they are actually usable (which it looks like they probably are). Most importantly for my personal self-confidence though, this means that my molecular technique is just fine, it was my understanding of the machines I was using that fell short. A deficit in knowledge is easier to correct and not such a blow to the ego.

RNA extractions and bioanalyzer results II

I have done a couple trials of RNA extractions to figure out what is going wrong (see my last post on rna extractions).

All these were done using liver tissue which had been dissected and snap frozen on dry ice:

I extracted the first two and the results were not great, so I talked with J~ and realized some mistakes (too much tissue, no Beta mercaptoethanol, undergrads using my pipettes...). Then I did the next three rows to correct for those possible errors. I also changed how I'm washing the mortars and pestels: use bleach, then rinse with molecular grade water, use RNase-away and rinse with molecular grade water, then cover with tin foil and dry at 95c, and cool before use.

I won't post the entire bioanalyzer results, but here is a summary of the results:
I also did an experiment where S~ and I each extracted RNA side-by-side to make sure that I wasn't missing a step or doing anything else weird. S~ didn't notice anything, but the results (both hers and mine) were well below 8.

As everything up to this point was only mildly successful if that (none above RIN=8), I decided to do a comparison between completely fresh tissue and that same tissue that had been frozen. I sacrificed an animal and divided the liver into two parts - on went straight on dry ice and the other went directly to the extraction. The one on ice remained there until I finished the first extraction ~1.5hours.
I also realized that I have a second protocol that doesn't require the use of columns at all (it just uses the RNAsolv reagent), so I decided to run these next to each other. I followed the same protocol as above in regard to fresh and frozen tissue. The RNAsolv method ended up with 4 tubes per treatment. I only ran 2 of each on the bioanalyzer.

Results:



The two RNAsolv with fresh tissue didn't work at all and the other two were quite poor. I won't be using that protocol again. As for the Fresh vs Frozen tissue with the kit, it looks like frozen is actually better, though the difference is small. At least I have gotten two 8's though, that's looking up.

Not really sure what to do next. Committee members have said to just keep going until it's right, regardless of how long it takes. But I don't have anything to change, and furthermore I don't feel like I've done anything different even between the earlier 6.6 and this recent 8. The lab tech has said that she didn't realize that she was doing anything different when she suddenly started getting high RINs though so maybe there is still hope for me.

In unrelated news, it turns out that placenta has one of the highest incidences of RNases (along with pancreas). That doesn't explain poor quality liver RNA of course, but is something for me to keep in mind when extracting my placentas in the future...