I just returned from a NESCent Catalysis meeting on Cephalopod Genomics. I was invited as a bioinformatics and genomics guy, and so I spent four days in North Carolina talking about the opportunities and challenges of sequencing cephalopods.
Cephalopods are a class of the molluscs, and include squid and octopus, as well as Nautilus. As part of the molluscs, they're within the broader Lophotrochozoa, a superphylum about which we know rather little, genomically speaking.
The critters are pretty strange and interesting. They're suspected to be quite intelligent, and are studied for their behavior. Octopus is well known for camouflage and mimicry (see Roger Hanlon's video). Architeuthis, the Giant Squid, is the very definition of charismatic MEGAfauna. Loligo squid is widely fished, and has a gigantic axon that is one of the classic models for neurophysiology. My friend Erich Schwarz, who was also at the meeting, said "this is as close as we're going to get to sequencing intelligent aliens!"
The goal of the meeting was to chart a course forward to sequencing cephalopod genomes. It was organized by Clifton Ragsdale, Laure Bonnaud, and Leonid Moroz, and attended by about 30 people. 20 or more of the attendees were members of the cephalopod community, while the remaining people came from sequencing centers or bioinformatics and genomics labs.
I had virtually no prior knowledge of cephs, and the first day -- on the known biology, phylogeny, and genomes -- was an eye opener. The smallest known ceph genomes, the pygmy squid Idiosepius, is around 2.1 Gb, and many of the genomes are 4-5 Gb, quite a bit larger than the human genome (at ~3 Gb). Phylogenetically, the cephs are a deeply diverged class and have about 700 species, separated by up to 300 million years or more -- as divergent as human and fish, roughly. Unlike vertebrates, they may evolve rather quickly at a sequence level, and to really use homology to connect sequences we may need to sequence a bunch of cephs at various strategically-chosen distances from each other.
The meeting concluded with a white paper writing session, and that will presumably be made available soon. In the meantime, here are some of the things that I found most interesting:
We ended up planning a sort of post-genomic consortium, as Eric Edsinger-Gonzales pointed out to me. So many people are already generating sequence (both genomic DNA and RNA) for cephs that the real question is how to organize, collaborate, and advance as a field, rather than how to start generating sequence.
The bioinformatics challenges for genomics on these critters are really big. Large genomes, presumably full of repeats; unculturable animals (meaning we can't inbreed quickly, and so are stuck with whatever polymorphism rate we get for a critter); divergent genomes and transcriptomes; and a really broad community scattered across 6 continents and studying between 4-10 species. I suspect that assembly and annotation of these genomes is going to be really challenging.
We (the bioinformaticians, collectively) vehemently argued for a liberal and explicit data-sharing policy. After a long morning discussion about pre-publication data release, we reached a few conclusions. First, so many people already have some sequence, and sequencing capacity is distributed so widely, that things like the Ft. Lauderdale agreement provide little leverage in pressuring people to release their data. Second, there has to be some positive incentive for people to release their data -- it's not enough to simply put in place protections for misuse of the data. Third, many biologists do not yet subscribe to the idea that data generation is relatively uninteresting compared to the analysis, and (given the way pubs and grants reward people for being "first") it's hard to blame them. Fourth, many biologists just don't see the point of making the data available outside the community. In response to all of these concerns, we put together a draft data sharing & repository proposal (here); if the community actually adopts it in the white paper, then it will be the most explicitly liberal small-community data sharing policy I've seen in biology. I have hopes.
Overall, I think it is increasingly hard to organize centralized or community funding for genome projects and databases. In cases where there isn't much funding or centralization, existing genomic resources supported by big sites such as Ensembl and UCSC can serve as good reference repositories. But I'm not sure what small communities such as the cephalopod community are going to do for the next few years. It looks like I'll be involved, so I'll let you know when I find out...
p.s. While not directly part of the cephalopod meeting, I had an interesting tweetstorm with Ewan Birney and Cameron Neylon this morning, where we discussed the draft community data sharing proposal that I'd posted. It ended with me typing up a section of the white paper in Google Docs while they kibitzed on various aspects of the writeup, live -- rather a unique experience for me :). Also of note, Casey Bergman posted links to Michael Eisen's data sharing license, the Batavia Open Genomic Data License, as well as his own musings. Cameron also shared the this page on principles of data sharing and the responsibilities of data providers and data users.
I came away from the discussion with Ewan and Cameron thinking that the cultural gulf between organismal and molecular biologists on the one side, and genomicists and bioinformaticians on the other side, was still really wide and hence hard to bridge. At least some significant part of this is driven by the culture of publication driven by the federo-Elseverien funding/citation/impact factor complex prevalent in biology.
Posted by Fabrizio Ghiselli on 2012-05-28 at 05:05.
I work on bivalve molluscs, and I understand 100% the challenge you're going to face. I think you're taking the right way, and I wish the same approach could be taken with bivalves one day... We need to know much more about Lophotrocozoans! Best wishes, Fabrizio