Parent-of-Orign Growth Effects Paper

J~ and I recently submitted my first first-author manuscript to Evolution. The reviews came back fairly positive, but there were some things that they would like changed. In total, I thought the reviews were really reasonable and I've already started collecting some extra data.

The main issue that they had was that I present evidence that some genes are properly imprinted in hamsters, some are polymorphic in their imprinting status within cross types, and some are expressed from two alleles in both hybrids. None of these patterns can explain POE growth as in all cases the two reciprocal hybrids show the same patterns. The reviewers were concerned that as I don't know the imprinting status of these alleles within each species, it is hard ("impossible" was the word they used) to interpret whether imprinting is breaking down in the hybrids. To address this apparent shortcoming, I'm planning on doing two things. First, making the writing clearer. I need to make it clear that if we see the same pattern of expression in hybrids, it cannot explain POE growth, regardless of whether the gene is imprinted in the parentals or not. While I think that should be sufficient, I am also planning on collecting more data in the form of a qPCR experiment. If I can show that the expression level is the same between parentals and hybrids, that is good evidence that the gene is not disrupted in the hybrids, and is therefore not imprinted in the parents. This should be easy(-ish) data to collect and so I'm going forth with some qPCR experiments.

There were other minor things, but all can be addressed in the text fairly easily. Hopefully I can turn this around and get it back to the reviewers soon.

DDIG

It's been a while since my last post. I've been working on a couple things recently that I should start writing about, but first, I should announce that I was awarded the NSF DDIG. The award money will go towards mapping the maternally inherited factors that result in POE overgrowth in dwarf hamsters.
The crossing scheme is this: make a small F1 (cam x sun) and then back cross those females to a campbelli male. See the figure.

I am focusing only maternally inherited factors because F1 hybrid males are sterile so I cannot do an F2 cross or a sun x F1 male backcross. As should be clear in the figure, the variation occurs due to recombination in the F1 female between P. campbelli and P. sungorus chromosomes, but that variation is only inherited through the mother, I cannot send variation through the father.

While it's kind of a bummer that I will be blind to anything that causes POE overgrowth when inherited from the paternal line, it's not the end of the world. We expect that maternally inherited things that are silenced are growth factors and their failure to be silenced is what causes overgrowth. In other words, we don't have an a priori expectation that paternally inherited factors are involved.

The grant money will support a RAD digest (Peterson et al. 2012) and Illumina sequencing in order to build the genetic map and map the causative factors. The shortcoming of using RADs is that they are anonymous markers - I don't know what genes are near them in the genome. In order to actually localize genes, I will do two things: (1) use a custom-designed array capture to sequence exons from the mapping panel and associate them with the anonymous markers  on the genetic map and (2) use RNAseq on the placentas of small and large backcross offspring to associate genes that are mis-expressed in those individuals with the genetic map.

I'm pretty excited about this whole approach as I think that it will give me the ability to get almost down to the gene level without fine-mapping. We'll see if I'm right or not, but it's pretty exciting that NSF is impressed enough to give me monies for it.

_______________________________________________
Peterson, B. K., J. N. Weber, E. H. Kay, H. S. Fisher, and H. E. Hoekstra. 2012. Double Digest RADseq: An Inexpensive Method for De Novo SNP Discovery and Genotyping in Model and Non-Model Species. PLOS ONE 7:e37135.

DDIG and the Government Shutdown

I'm currently writing my DDIG (Doctoral Dissertation Improvement Grant) which will be submitted to NSF. It's going pretty slowly, but hopefully I'll have something worthwhile here shortly. Fortunately for me, the House Republicans were kind enough to shut the government down and give me a couple extra days to write. Thanks John Boehner! Thanks Ted Cruz! You guys are the best!

New Hamsters!

We just received a new strain of Siberian dwarf hamsters. They are an inbred line from Bruce Goldman at UMASS Amherst. They are reported to change color reliably and, as they are a different collection from the stocks that we've been using so far, they should be genetically distinct from ours to some degree.

I'm hoping to use them to assay imprinting within P. sungorus if they are different enough. Of the projects that I've done so far, the main shortcoming is that I do not know which genes are imprinted in hamsters and which are not. I have to guess based on the genes that are imprinted in house mice, deer mice, and humans. This is a problem because we know that different species imprint different genes. For example, deer mice do not imprint Mash2 while house mice do. Of the genes I surveyed in our original strains of dwarf hamster, there were three that appeared to not be imprinted, but that may instead be due to imprinting breaking down in both reciprocal hybrids. Using this new strain of P. sungorus I should be able to determine which genes are actually imprinted within that species.

This will of course depend on whether there are an appreciable number of nucleotide differences between the two strains. In order to determine whether an allele is imprinted, I have to find at least one diagnostic SNP within the expressed region. If these two strains are very closely related, they may not have such diagnostic SNPs. However, I think that they likely will as between P. campbelli and P. sungorus there are 3-6 diagnostic sites per kilobase and those two are relatively recently diverged. The standing polymorphism within one species should be less, but likely still high enough for there to be at least 1 SNP in each exon. I am assuming that these two independent collections will have sampled different subsets of the standing species-level polymorphism and so will have enough differences to answer my question.

candidate imprinted gene sequencing experiment

I have a paper that I'm working on in which I have sequenced the cDNA from 8 candidate imprinted genes in both species and the hybrids. I did this in order to determine whether imprinting is breaking down at these loci. I found that there is no signature of imprinting breakdown in 7 of these genes (the eighth had no diagnostic differences) in 6 hybrids (3 P. campbelli x P. sungorus and 3 P. sungorus x P. campbelli). However, we have since realized that there may be sex-specific issues and as the original 6 ended up being 4 males and 2 females (we had no way to sex-type the offspring at the time) we have decided to repeat this experiment with 3 males and 3 females.

Fortunately I have all the RNA extracted for the next-gen project and so all I had to do was synthesize cDNA and amplify then sequence the genes of interest. I'll update with results.

So far 6 of the 7 candidates that have diagnostic fixed differences have been sequenced. The results remain the same as before: no breakdown of imprinting in one hybrid: there are some genes where breakdown occurs in both hybrids and some where imprinting is correctly maintained. Furthermore, there are no differences between the two sexes in the pattern of expression.

The 7th gene, Peg3, is not amplifying and so I can't sequence it. It is an important gene to assay because it is one of the main partners in the deer mice that cause parent-of-origin growth (Vrana et al, 2000). It is frustrating that it won't amplify reliably as samples I ran a year ago were great. The primers do not include an intron and so the first thing I will do is to try the primers on gDNA instead of cDNA - maybe the gene is not expressed highly in the placenta of hamsters and that is why the PCR won't work.

update 7-15-13: The PCR with gDNA worked fine, not a brilliantly blazing as the first time I did it, but not bad. This leaves me with the question of whether the original PCRs had gDNA contamination with the RNA. I think that this is not the case as Peg3 showed imprinted expression in those samples. If gDNA was a contaminate, the sequences would have been heterozygous and shown what I would have interpreted as biallelic expression. The only way imprinted expression could be found is if the gene is truly imprinted and there was no gDNA contamination.

Now the problem remains of how to get the Peg3 primers to work on cDNA like they did before. I have increased the extension time by 20 seconds in case the full 1000bps weren't being copied and I increased the cycle number from 35 to 40. Although I'm a little skeptical of amping anything 40 cycles, if there are faint bands there, 40 cycles should make them show up.

update 7-16-13: All the cDNA bands are really faint if they're there at all. I'm pretty sure the positive is present, but it's kinda shitty all the way around.  E~ suggested that I should do a comparison between the gDNA of the original sequences and the gDNA of the one's I'm struggling with now to tell if there has been a problem with the primers (all gDNA's fail) or perhaps a SNP in the priming site (new gDNAs fail, old ones work). Although this is going against the Prime Directive (Science's General Order no. 1: only change one variable at a time) I am using new Taq too. This is partly because we are out of the old, and partly because I think the old may not be working. Here is my  predictions/interpretation table:

IF:                                               Then:
Both old and new gDNA fail     primers are bad, reorder primers
Old gDNA works, new fails      SNP in the priming site, redesign primers
Old gDNA fails, new works      WTF, I have no idea what this would mean, I hope it doesn't happen...
Both old and new gDNA work  old Taq was bad, the new is good, re-try with cDNA

update 7-30-13: here is the gel:

Land 1 is the ladder, lanes 2-7 are the samples, the first three are the old gDNA, the second three are the new gDNA, lanes 8 and 9 are the positive and negative control, lane 10 is another ladder. 

All the gDNAs worked, so I will assume that the issue was bad Taq and tomorrow I will run out the cDNA with the good Taq.

update 8-1-13: Didn't work. The old cDNA still doesn't amplify. Try making new cDNA

update 8-14-13: I figured it out! I had cleaned the cDNA in the previous trials and since I had started with so little, I lost most of it. This time I did not clean the cDNA and have beautiful bands. Now for sequencing!

Lane 1 is the ladder, 2-18 are Peg3 from cDNA, 19 is the positive control and 20 is the negative control.

Such a ton of work for so simple a problem. I'm just glad it's working.

----------------------------------------------------------------------------------------------------------------------------------
Vrana, P. B., J. A. Fossella, P. Matteson, T. del Rio, M. J. O'Neill, and S. M. Tilghman. 2000. Genetic and epigenetic incompatibilities underlie hybrid dysgenesis in Peromyscus. Nat Genet 25:120–124.

RNAseq Libraries sent!

I finished pooling my samples (lesson learned: don't use small volumes it evaporates as you pipette) a couple weeks ago and they've been sitting in the fridge ready to ship off just waiting for the finances to come through with the University of Utah. All the paperwork was finalized yesterday and we just missed the shipping time (dropped the package off at 4:05 for a 4pm pickup). We had to pick it back up and hold onto it until today. We'll add some more dry ice and send it off this afternoon.

It has been a ton of work to get that 30ul ready for shipping. Now we cross our fingers and hope that it comes out all right. In the mean time, I think I'll have a beer.

Evolution Conference 2013 - Tuesday

Many more talks today, here's some notes on the ones that stood out:

Jennifer Kovacs told us about a cool ability of the endosymbionts of aphids and how they defend against ladybugs. When a female ladybug eats the infected aphids, she suffers no immediate consequences, but her offspring show a reduced ability to successfully pupate. Those that do pupate are on average larger than normal which may be due to a “filter” effect where the smaller, weaker ones were weeded out by the bacteria. As aphids reproduce clonally, the relatedness between two aphids on the same leaf is 1. This means there is strong selection to stop the adult ladybug’s offspring from consuming other aphid clones. The only hole in the story so far is that no one has been able to show that the aphid’s bacteria survives the adult ladybug’s stomach and gets into her ovaries. Somehow it must get into her eggs, but they haven’t been able to show that yet. Still a really cool story of host defense. 

Taichi’s talk was fun too. He’s looking at Bergman’s Rule (body size increases with latitude) and how it correlates with the gut microbiome. He’s showed that especially in the east coast there is a strong correlation between composition of microbes and latitude that also correlates with average body size in house mice. There is less of a pattern on the west coast, but that is likely because the colonization was much more recent and there has been subsequent gene flow from Mus spretus that may be confounding the pattern. 

Jamie Zuniga-Vega talked about superfetation in live-bearing fish. Superfetation is where one female carries many broods at different developmental timepoints concurrently. There are three main hypotheses for superfetation which are: (1)  it lowers the peak cost of reproduction - at any given time, female must invest less per unit time. (2) It results from morphological constraint - many offspring take up a lot of space forcing the female to not be hydrodynamic anymore, if the offspring are at different stages, they take up less space. And (3) it compensates for high adult mortality. If there is a high probablility of adult mortality, high fecundity may be able to comensate. Thus superfetation may increase rate of offspring production rate. They found that hypothesis 2 is best model for two species where fast water causes higher rates of superfetation, but the third species doesn't have anything that shows preference of one hypothesis over another.

Keenan Morrison then talked about anamniotic eggs and how they may predispose species for the evolution of matrotrophy.  Amniotic eggs are impermeable and cannot absorb nutrients, but anamniotic eggs can assimilate resources across the “shell.” As these eggs already can accept nutrients from the environment, it is easy to make the next step to accepting nutrients from the mother’s uterus.

David Reznick followed the previous two talks and discussed how placental or matrotrophic fish should have greater male-female conflict. Furthermore he argued that placentas should correlate with long gonopodium (penises), small body size, lack of courtship, and sneaker mating strategies, non-placentals on the other hand should show strong sexual dimorphism, elaborate mating rituals etc. I was unclear on his logic as to why this is true (can’t placentals have pre-mating sexual selection as well as post-mating conflict?) and am looking forward to reading a paper he will hopefully publish ont eh topic (I felt better about my confusion as Doug was also unsure of his logic here). Surprisingly, he has found that there is an elevated speciation rate in non-placentas compared to placental fish. This is opposite of what I would have expected given the opportunity for maternal-offspring conflict in placental species. The species that do not have placentas instead show elaborate mating rituals and are characterized by strong premating sexual selection. This may then go hand-in-hand with the observation that birds, where strong premating sexual selection/sexual dimorphism is common, speciate twice as fast as mammals, where developmental conflict is more pervasive. Reznick also found that there is no support for “adaptive hypotheses” of the origin of placentas, but I’ll need to see his paper to remember/understand why he thinks so. Hopefully the paper with come out soon. 

I was lost with Turelli’s talk. He went fast and had a ton of words on his slides. It sounded impressive, but I couldn’t follow anything he was saying. 

Rob Unckless’ talk was a great conclusion for the day. He talked about the drive system in Drosophila affinis. He and I had spent many hours talking about this system back in the Jaenike lab and creating recursion equations to model it. the equations we made then were comparatively quite simple - one driver, one suppressor. What he talked about today was how this system may maintain polymorphism of the Y chromosome given many drivers and many suppressors. Apparently he had to make a perl script to generate the recursion equations as they got so complicated. I am going to invite him out for a talk in the spring if I can manage it. 

The last talk of the evening was Jack Sullivan’s SSB presidential address. It was really fun and he showed great pictures of bacula and was quite entertaining. Much of his talk was work done by J~ either as his masters student or recently in collaboration. He ended with the statement that you should always treat your students well so that you can exploit them years into the future. 

I got to hang out with Rob and Yaniv, Erica, Mike Shapiro, Yasir, Daniel Matute and a bunch of others in the evening which was a blast.  All in all it was a superb conference.