下一个被测定的基因组是哪个?

[编者的话]

人，小鼠、果蝇，这些模式生物的基因组已被测定。那么下一个被测定的基因组是什么呢？下面这篇文章谈及了那些基因组应该被考虑，以及选择的标准。应该被考虑到因素应该至少包括：科学界对测定这个基因组的热情高吗？关于这个生物的现有资源多吗，相关技术是否成熟？该生物适合用来做实验吗？基因组大？测定之后的实际意义很重大吗，比方说可以给人类疾病的研究提供信息和资源…………

by Tabitha M. Powledg

Human, mouse, fruit fly - those genomes and a few others are nearly out of the way. What next? Genome experts are in the midst of an elaborate effort to figure out not just which genomes should get priority, but which of their features should count in setting those priorities.

How to set priorities, rather than which genomes to pursue next, was the ostensible topic of a gathering of scientists convened last week by the National Human Genome Research Institute (NHGRI). The question before it, according to NHGRI director Francis Collins: "How do we avoid greasing the squeakiest wheels?"

The assembled researchers could not resist squeaking, of course, and by meeting's end they were being urged to do their lobbying formally, in white papers making the case for their pet organisms.

Still, some choices are so obvious they look like a done deal. Apes top the list, especially our closest kin the chimpanzees. (This would be the common chimp from US zoos and research institutions, rather than the sexy, amiable, and rarer bonobo.) During the course of the meeting other candidates began to seem increasingly logical. The rhesus macaque, for example, is plentiful and has long been a model organism, both in studies of cognition and perception and for medical research, including AIDS and vaccine development. Moreover, the range of experimentation permitted on the rhesus macaque is wider than what is allowable with apes.

The primate shopping list is potentially a long one. Certain New World monkeys have interesting family lives and interestingly jumbled genomes. Marmosets are easy to breed. The mouse lemur breeds quickly and prolifically, and is of evolutionary interest because the primate radiation began with its appearance. Neuroscientists might favor rhesus monkeys (not macaques), which have human-like moods and mood disorders with heritable, clear, neurobiological correlates.

There are nearly 5,000 other mammals out there too, and the National Cancer Institute's Stephen O'Brien has long argued the case for them. He pointed out that all mammals being sequenced at present belong to Clade III, just one of the four mammalian lineages that diverged some 80 million years ago. So he urged that representatives from the other groups should be sequenced as well.

Sequences from Clade I (elephants and aardvarks) and Clade II (armadillos and anteaters) seem like long shots, and so at this point do marsupials and monotremes. But animals from O'Brien's Clade IV are likely to be popular sequencing subjects. This group includes livestock of major commercial importance - pig, cow, sheep, horse, and others - and the "companion animals," cats and dogs.

Grassroots demand from pet owners and the consequent popularity in Congress make dog and cat genome projects likely, but that doesn't mean there aren't also sound scientific reasons to sequence them. For one thing, the sequences would be informed by the immense databases of medical information assembled by vets, noted Elaine Ostrander of the Fred Hutchinson Cancer Research Center in Seattle, who is a passionate advocate for a dog genome project.

Dogs are good models for human disorders - examples are deafness and heart disease - and important pharmacological models for transplant drugs. Since they are astoundingly diverse morphologically but genetically quite homogeneous, dogs could provide insight into developmental genomics. Having always been bred for their behavioral phenotypes, dogs also intrigue behavioral geneticists. Ostrander says her freezer is stocked with tissue from a dog family with obsessive-compulsive disorder in which the animals respond to drug treatment just as humans do. She has maps in hand, BACs in hand, and is ready to go. "We've mapped hundreds of genes," she said at the meeting. "An investment of just 2x genome coverage would really pay off."

Undoubtedly it strengthens an organism's case to be warm and fuzzy, but less charismatic creatures also have strong advocates and strong arguments. Wouldn't we like to know how to regenerate damaged tissue and organs? Susan Bryant (a developmental and cell biologist at the University of California at Irvine, and also dean of the School of Biological Sciences there) argued for the salamander Axolotl, the only vertebrate that regenerates well, and Laura Landweber, ecologist and evolutionary biologist at Princeton, urged attention to protists. Tetrahymena, she pointed out, contains many human homologues not found in yeast.

In a follow-up email, she noted: "If you look at the diversity of eukaryotes, most of their approximately 2 billion years of evolution is protists, with plants/animals/fungi just occupying the few crowning branches of the eukaryotic tree. Therefore, if you want to learn about eukaryotic genetic diversity ... you have to look at protists. This also allows you to put other metazoan genomes, like C. elegans, D. melanogaster, or ours, in the context of the big tree of life's three domains (Eukarya, Bacteria, Archaea), filling in the countless lineages between prokaryotes and ourselves."

Pleas for attention to the plant kingdom and microbes got short shrift. "This is one of the boundary questions of which agency funds what. There are other sources of funds for other projects," Collins said. NIH has not been supporting sequencing of plants, which are a better fit at the National Science Foundation and the US Department of Agriculture, he argued.

"We need to be able to make a case that there's a relationship with human health," Collins said. To which Mitchell Sogin of the Marine Biological Laboratory at Woods Hole, Massachusetts, responded, "Looking at organisms that live inside of us - this is not irrelevant to health."

Generic issues were plentiful too. How to judge whether genomic information from a particular organism would be profitable? How to decide between organisms that are useful experimentally and those whose genomes might yield information about gene regulation, or mutation rate, or developmental timing, or evolutionary innovation in function? Must we have complete, finished sequences? Or will partial or regional or draft sequences do, especially for some kinds of information and some organisms? How can other fields make use of genomic information, especially neurobiology? Will costs continue to drop, permitting sequencing of additional creatures? What new technologies and resources can scientists expect - or hope for? How can sequencing plans be coordinated internationally to prevent duplication of effort and unproductive competition? What kind of information does bioinformatics need in order to refine tools for genome analysis, and can genome experts provide it?

From the discussion, meeting cochairs Robert Horvitz of the Massachusetts Institute of Technology and David Botstein of Stanford University School of Medicine spliced together two tentative lists of criteria, general and specific, that will probably figure in decisions about which genomes to sequence when. Botstein noted that these issues should be addressed in any proposal to sequence a particular genome - although he hastened to add that the list should not be regarded as just another NIH form to be submitted with a request for funding.

Horvitz said, "Remember nothing on this list is a categorical requirement. It could be that there's a genome that fits none of these but still should be sequenced." The list, he declared, was a list of examples, and not all of them need apply to every organism.

Among the general considerations:

• How big is the research community that could put sequence information to work, and is it enthusiastic about having the sequence?

• What resources already exist for the organism? For instance, how much genetic information is already available? Do BAC libraries exist already? Are technological methods in place, such as gene transfer? "Obviously, your stuff is going to get sequenced sooner if BAC libraries exist, and those shouldn't be left to the genome centers," said Botstein.

• How big is the genome? The smaller the better, because sequencing costs are linear. This requirement is the biggest strike against learning about regeneration from Axolotl's genome, which is 10 times larger than ours. It also prompted the Whitehead Institute's ubersequencer Eric Lander to point out that sequencers need much better ways of dealing with large genomes.

• Is the organism suitable for experimentation - easy to house, raise, breed, study? If it isn't, this need not be disqualifying, Horvitz pointed out. The puffer fish Fugu, for example, is not a good model organism, but its junk-free genome makes it so attractive that the Department of Energy began sequencing it last year.

• And, oh yes, relevance to human health. Most of the funding will, after all, come from NIH,

On the list of specific rationales:

• Will the organism's genome inform the human sequence, inform the sequences of other model organisms, and/or and serve in some way to connect the two?

• Will the sequence develop or support genomics for a model organism new for that field, although not necessarily a new model organism? Examples here are the rhesus macaque and, for developmental biology, the sea urchin. Lander warned researchers not to regard this criterion as an invitation to extend the list of model organisms. "It's information we want, not new models," he said.

• Would sequence information bolster knowledge about basic biology, especially developmental and neurobiological?

• Would sequence information illuminate the course of evolution, the relationships among species, and the process of functional innovation?

• Is the organism a good surrogate for human experimentation? This could include not just disease models and suitability for drug testing, but perhaps other topics, even if highly specialized. Here is a potential starring role for the most offbeat candidate proposed at the meeting: Spermophilus tridecemlineatus, the 13-lined ground squirrel, whose torpor during hibernation might yield information valuable to transplant patients and their surgeons.

It is not yet clear how NIH will go about making decisions about specific organisms. Collins said he hoped the peer review process would evaluate the big picture, not just judge individual proposals.

Collins also emphasized that the agency did not want to ram a project down the throat of a sequencing center. He urged researchers interested in particular sequencing projects to seek alliances with the big sequencing centers and rev up their enthusiasm. NHGRI, he said, would probably support travel money and pilot projects. The aim: "We'd like this to bubble up from the scientific community."