Hello all - 

We are going forward with sequencing of the Great Star Coral (Montastrea cavernosa) genome. Genomic DNA has been extracted, passed all quality controls, and is now waiting in queue to be sequenced using the third-generation sequencing technology, PacBio RS II. This technology sequences single DNA molecules and produces extremely long reads, in excess of 10,000 DNA letters (for comparison, the most commonly used technology for sequencing human genomes, Illumina HiSeq, can only go up to 200 letters per read). Such long reads will make the genome assembly much easier.

The samples also have been transferred to our collaborators at the Ohio State University who are preparing Hi-C samples for sequencing. With these, we can assemble the whole chromosomes without gaps. Again, for comparison, new animal genomes that are being published today are collections of thousands of disconnected fragments.

So, overall, this coral genome with be a great step forward in genomic work. We are very excited about it and hope to make it a demo case "this is how you are supposed to sequence new genomes in late 2010s", for other people to follow.

cheers

Misha

Hello all - we are more than halfway to the $3,200 mark, which will allow us to produce the first draft of the new coral genome! This would be really exciting.

With every $50 we are adding another coral individual to analyze present-day genetic diversity in corals - a small but eminently worthwhile project. If we make it to $2,000, this would give us 40 individuals - a possibility to compare two populations in one coral species in terms of shared population history and ongoing genetic exchange. This could be a paper all by itself! 

cheers

Mikhail

Hello all - 

I was asked these two important questions by one of our coral research colleagues, and I thought I'd post my answers here as well:

Q: How can you do this unless you have samples of corals from throughout the last glacial cycle? 

A: You can reconstruct past population sizes without any ancient coral samples - that’s the whole point. 

Two methods have been recently developed that can do this based on present-day genetic variation. Both are based on coalescent theory and have been extensively tested on human populations, for which we kind of know what happened in the past. 

The method called dadi (diffusion approximation for demographic inference) examines allele frequencies at thousands of point mutations (called "single-nucleotide polymorphisms", or SNPs)  in populations and is based on two notions: - mutations in the genome accumulate at a rate proportional to the population size;  - more recent mutations tend to be found at lower frequencies than older mutations. Coalescent theory provides equations to link these together and make it possible to fit a model to the observed mutation frequency data incorporating population size changes through time. Advantage of this method is that you don't need to sequence the whole genome and it does not matter where those mutations are in it; although you do have to genotype 15-20 individuals per population at 10,000+ SNPs. Our 2bRAD method is just the tool for that. 

The other method - PSMC, pairwise sequential markovian coalescent - requires just a single individual genome sequence (I know, it sounds like magic but it works). It makes use of the fact that average distance between neighboring mutations in the genome decreases when populations expand, and increases when they contract.  So it basically fits a time-resolved model to match the distribution of distances between mutations. One great advantage of this method is that you don’t need to specify what sort of pop size changes you expect to find - unlike dadi, which requires a pre-specified sequence of changes and the model finds the best-fitting time and size parameters for it. The difficulty with PSMC is that you need a whole-genome sequence, which could be fragmented but still must be separated into individual homologous chromosomes (“phased” is the term). This is why we propose to use Hi-C method that provides such information.

Q: And if you could reconstruct that, how would that help adaption to conditions outside that range?

A: We anticipate that for adaptation, the rates of environmental change rather than absolute temperature reached would have the most impact. Corals can live in the Arabian Sea at brutally high temperatures, so we don’t seem to be nearing the absolute physiological temperature cap for them yet; the question is, is the present warming too fast for them to keep pace? There were some very rapid temperature changes in the Northern hemisphere since last glacial maximum, most notably the Younger Dryas, about 11,500 year ago: the temperature in Greenland increased by 7oC or more in just a few decades. Did this coincide with the collapse of Caribbean coral populations? If not, that would be good news. 

Of course this would not give us a complete picture for the present situation - we have other climate factors and human influences to put into the equation as well. But at least we would know that, even though the climate will inevitably continue to warm in the next few decades, region-wide ecological extinction of corals is not inevitable, and so our efforts to save reefs on the local scale can actually make a difference. 

cheers

Misha

Hello all,

Many thanks to our close friends and family who made the vast majority of donations thus far. Still, we are even more grateful to those precious few who made their decision to donate without actually knowing any of us personally. Thus far we have only one or two of these brave people, demonstrating that they care about coral reefs and believe in us. We very much hope there are more of your kind out there! 18 days to go...

I thought I'd post some additional information with this update, for those who might be interested in digging into details of how we do our research and why. Recently Jorge Salazar of the Texas Advanced Computer Center (TACC) published a story on our coral genomics project making use of TACC supercomputer power: https://www.tacc.utexas.edu/-/supercomputing-coral-s-race-to-beat-heat . It goes into explaining many details of our work published this summer in the journal Science ( http://www.sciencemag.org/content/348/6242/1460, http://nyti.ms/1P6H8uE), and even includes a podcast (in which I don't sound half as bad as I was sure I would).

For those who would like to see more about coral decline around the world, here is a very good compilation of available data thus far, put together by coral reef ecologist John Bruno (@JohnFBruno) of the University of North Carolina at Chapel Hill: http://theseamonster.net/2013/09/what-we-know-about-coral-loss/ . The take-home summary: As far as we can tell, the coral cover average of most regions is well below 20% and often closer to 10%. Therefore, I think we have lost at least 50% of living coral cover and more likely 60 or 70%.  And in some regions quite possibly as much as 80-90%.  And this is worldwide, not just the Caribbean. Pretty sobering.

cheers

Misha