Our understanding of vertebrate genetics and development has been helped immensely by the ability to target genes for deletion in the mouse, a technique called knockout. Once one copy of a gene is eliminated, animals can be bred that lack the gene entirely, allowing the gene's role in development, behavior, and health to be assessed in an organism that's closely related to humans. Many of the genes identified as being important in other organisms or via biochemistry have since been knocked out. But this technique has its limits, not the least of which is that there's a tendency to only knockout the genes that we expect will be interesting, and labs wind up racing to be the first to knock out the most interesting genes.
Last week's edition of Science takes a look at the state of the art in mouse knockouts. One article looks at some decisions involved in the NIH's program to create a publicly accessible collection of knockouts in every single gene of the mouse genome. The state of Texas had funded a combined public/private consortium that hoped to get a big slice of the NIH work, but seems to have missed out on the money. An official explanation of why hasn't been released yet, but speculation suggests that concerns exist about both the technology used and the accessibility of the mice produced by the Texas group. Meanwhile, in China, Fudan University hopes to get in on the NIH action via a partnership with Yale University. Instead of targeted knockouts, these researchers are getting transposons (mobile genetic elements) to hop around the mouse genome and searching for cases where they have hopped into genes. Although not as directed as targeted knockouts, the ease of generating transposon hops allows large numbers to be screened, potentially making up for a lack of efficiency.
The final piece looks at why the NIH effort and a parallel project in Europe seem to be necessary: rampant inefficiency. It quotes a researcher that has attempted to obtain a number of mice that have been generated by others:
"Once I requested a mouse, and the guy wanted everyone from himself to his grandmother to be a co-author on everything we published with that mouse," says Aguzzi. "It was like scientific prostitution." Another time, he says, a researcher promised him a mouse but took more than a year to deliver: "[The investigator] should have just said his cat ate it; it would have saved us a lot of trouble."
But some mutations are far too easy to obtain. Of the over 11,000 genes knocked out, over 700 had been targeted at least three times by separate research groups, and one had been hit 11 times. Over a quarter of the mice aren't in publicly accessible collections.
Even assuming better public access can be achieved, a number of significant problems will remain. Several companies have set up mouse knockout services, and have patented the specific deletions they offer. Aside from IP issues, there are biological complications. Animals with a developmental phenotype that survives to birth will be easy to identify; behavioral defects or embryonic death can be much more challenging to characterize, and it's unclear how well that will work out on the scales of these projects. There's also the question of technique. Knockouts can be complete deletions, can have a marker inserted in place of the gene, and/or can be set up to be deleted in specific tissues or have the deletion induced by a specific drug treatment. Different programs are focusing on different techniques. Meanwhile, an increasing number of researchers are focusing on the classical genetic approach of randomly mutating mice and looking for interesting phenotypes before going back and identifying the genes (an effort I've been involved in). Each technique has its benefits, so I expect that integrating all of these mouse sources and generating a consensus on how to push forward in an organized way will be extremely challenging.