Sex-typing pure species embryos and the RNAseq experiment

Now that I have the gDNA, I've set up the PCR to test for sex. The PCR appears to have worked well. So far for P. campbelli (lanes 1-22) I have 10 males from 3 crosses and 12 females from 6 crosses. From P. sungorus (lanes 23-40) I have 10 males from 5 crosses and 8 females from 5 crosses. This should work well for the RNAseq experiment as I want 5 of each gender of each species. However, I will extract a few more P. campbelli males as it would be nice to have 5 litters instead of 3 that way every individual represents a different replicate cross.

Below is the gel. Positive bands mean that individual is a male, whereas negative bands indicate that the PCR failed and I am interpreting that to be a female. Alternatively failed reactions could be males and failed for some other reason - it's hard to tell the difference between these two alternatives. I am assuming that there will be some non-specific amplification in the females, so a smear indicates that the reaction would have worked had there been a Y chromosome (rather than failed outright) and is thus an actual female, whereas a completely blank lane indicates a likely failed reaction. I'll only use "females" who showed some non-specific binding rather than the "females" where there is no product in the well at all (fortunately, all the "females" have a smear, none failed completely).

As I have alluded to it numerous times, here is the outline of the experiment and the motivation for the sex-typing:

I want to compare the imprinting (and more generally, the entire expression profile) in Phodopus hybrids. As the hybrids show drastic parent-of-origin dependent growth differences, there should be some pretty striking parent-of-origin dependent differences in gene regulation (likely a species-specific breakdown of imprinting). The general approach is to compare the profile of the reciprocal hybrids with each other and then to expression of the parents. The between-hybrid comparison can tell me about imprinting and whether it is disrupted in any specific manner as well as whether there are any transcriptome-wide changes in gene expression. The hybrid-parent comparison can tell me about overall level of expression and if it is in fact different between hybrids, then I can use the parent species as a control for "normal" levels.

We are unsure of whether there will be sex-specific expression differences though and to account for that - there are genes that are predicted to show differences between the sexes during development - I will use 5 of each sex from each cross type. This results in 40 samples, 5 biological replicates per genotype, which will hopefully give me enough power to ask the questions I'm interested in.

The sex-typing from above will help me choose which samples will be the ones that I prepare.

gDNA extractions

The one thing that my RNAseq experiment (more about that later) needs is sex-specific data. However, I haven't been able to determine the sex of pure spp embryos as my sexing technique relies on heterozygosity between spp-specific fixed differences (thus it only works in hybrids). However, I have recently designed primers for SRY, the mammalian male-determining gene, and can now test the pure spp offspring.

This requires that I have gDNA from all my 120-some pure spp embryos. I dislike gDNA extraction because they never work quite like they should. We use a Machery-Nagel Nucleic acid and protein purification kit, which is a column based kit. This means that if you don't fully digest the tissue the column membrane will clog and turn an hour long protocol in to a six hour protocol. 

I figured the reason it clogged is because when I add the protinase K I make a master mix of buffer and enzyme which sits for an hour or so as I collect the individual tissues. This may cause it to self-digest. So today I added the protinase K last to each sample after I added the tissue. I hoped that this would keep the enzyme's reactivity high. It appeared to work as all the digested tissues had little if any chunks of tissue left. However once I started the spin it was clear that something went wrong. Normally 20-30% clog, this time every single one was clogged. 

The only thing to do then is spin for longer and faster. This works, but takes tons of time and can get really frustrating. This time was actually quicker than most though. 

To get around this problem I have tried a number of things:

--varying the incubation period - overnight vs 3-5 hours

--assuring that the protinase K is not degraded

--pass the digested tissue through a syringe

--centrifuge the digested tissue and only use the supernate

Nothing alleviated the problem 100%. Perhaps a combination of fresh protinase, and the centrifuge would do...

Also it might be worth trying grinding the tissue on liquid Nitrogen... 

#################

The other thing that I have done today is set up a PCR checking the sex of the hybrids using my SRY primers. I assayed their sex earlier with ZFX and this is a quick check to make sure that ZFX and SRY agree. I'll post that gel tomorrow once I run it out. In fact, I only did the first 94 samples of nearly 120 as that's the capacity of a PCR machine. I'll do the rest tomorrow. 

I had a horrible time designing the SRY primers last summer, that's what led to the ZFX (which were also kind of a bitch to design). But last month J~ told me to take a week and see if I could get them working and that's all it took. I had to use every SRY sequence in genbank for cricetids, align them, build a tree, infer where my hamsters are and predict what the best sequences would be. Then, after a bunch of PCR optimizing I sequenced the messy product, and found redesigned primers based on actual hamster sequence. 

Here are the primer sequences I ended up with: 

>SRY5F

TGAATGCATTTATGGTGTGG

>SRY6R

AAGGTCTTCAGTCTCTGTGCTTC

the product is 166pb

Tm=60

Te=20sec

Cycles=35

Here is a temperature gradient from 54 to 60 showing the male band at around 166bps and the shorter primer dimers for the females. For some reason Blogger chopped off the labels I put on the ladder: the bottommost, quite faint band is 100bps, the next up is 200bps. My bands fall out right in between. The X-ed out parts are the same temp gradient but a different reverse primer that didn't work so well - it picked up non-specific stuff in the females and was just not nearly so robust. The four bands up top marked with "+"s are the same four DNA samples with a different gene, cytb, for positive controls.

Why am I writing a blog?

I dunno.

I went to a talk by Rosie Redfield at the Evolution meeting in Ontario last month where, surprise surprise, she told us all that we should start blogging. I doubt that's the real reason I'm doing this though.

It's probably more that she scared the shit out of me in her talk. I don't want to make the same mistakes that they did. Maybe this will help, maybe it won't. In all likelihood this blog will last the entire week, which I anticipate will total 2 posts and I'll forget all about it.

In any case, her talk was about the arsenic-based life paper that came out a while ago and all the fails that happened in order to get that paper published. She gives a great talk. She pointed out each step in the research that went wrong with great humor and the audience and myself laughed throughout. We laughed at the idiocy of those poor researchers who didn't do good science, we laughed at all the mistakes they made and how they should have known better.

And yet, I found myself thinking- even while laughing- yikes, I would have done that  ...  I would have fallen into the same trap  ...  I would have made that exact same mistake and never realized. Never realized that is, until Rosie herself ridiculed me to the audience at the next Evolution conference.

Honestly though, it wouldn't have been that bad for me, most of the mistakes she told us about I was right on board with - how on earth could someone do whatever-it-was and think it was ok? But then again, there was more than one mistake she described where it could have just as easily been me the one being roasted.

That's what scared me so bad - they were mistakes that I wouldn't even realize I was making.

So why blog?

I think it's because the biggest problem that happened with the arsenic research is that the lead author "fell in love with her hypothesis." What an easy trap to fall into. Rosie explained that in science there are the experiments you do when you think you are right, and the experiments you do when you honestly want to know if you are wrong.

I hope that by writing this out, by making it public, leaving no shadows to hide in, and no excuses to stand behind, I will come to terms with my loving hypotheses. I will do the experiments that will tell me I'm wrong.
And maybe someone will tell me if I'm making an unwitting mistake.