Friday, November 14, 2008

Odorprints on a Wanted Poster



In Ray Bradbury’s sci-fi novel Fahrenheit 451, the fire department uses a hellish Mechanical Hound to track a person down based on his individually unique body odor. The “sensitive capillary hairs in the nylon-brushed nostrils” of the Hound pick up the scent. Its electronic brain is programmed to recognize ten thousand individual BO profiles. After cornering its prey the Hound dispatches it with a lethal injection from a retractable fang.

I thought of the Mechanical Hound the other day when I spotted a juicy headline on Drudge: ‘Odorprinting’ will identify people. I clicked on the item with my usual mix of emotions. Would this be a story about real technology or some type of National Enquirer nonsense?

The linked report from the Telegraph quotes Jae Kwak, a scientist at the Monell Chemical Senses Center in Philadelphia. According to Kwak, genetically-based body odors are as unique as fingerprints; he thinks it may be possible to build devices that identify people based on their BO.

Intriguing. But in world of science-by-press-release it pays to be skeptical of media stories. Kwak could be a quack. Far better to examine the actual study that brought him to the Telegraph’s attention: in this case a paper in the open-access, online journal PLoS ONE.

The needle on my Bogosity Meter twitched as soon as I began to read: the experiment has nothing to do with humans—it’s about two strains of highly inbred mice.

I’m very familiar with these mice: I studied them years ago when I was on the faculty of the Monell Center. Originally bred for cancer research, the strains are genetically identical except for a set of genes that controls immune response. These so-called MHC genes are associated with a strain-specific scent. Mice can smell the difference and prefer to mate with mice of the other strain (a case of “opposites attract”). In addition, the mousey BO is distinct enough that humans can smell the difference between the two strains. (An experimental result of mine that I published here.)

It is widely believed that diet alters body odor. If so, can diet obscure MHC-linked BO differences? This is the question Kwak and his colleagues set out to answer using the two mouse strains and two kinds of commercially produced mouse chow. They trained “sensor” mice to recognize the scent of urine from mice with a specific combination of MHC type and diet. Then they let them choose between mice differing in various combinations of genes and diet. The result? The sensor mice failed to pick out the MHC-linked scent. In a direct match-up, diet-related BO overwhelmed the differences in genetically-based BO.

Other mice were trained to recognize a particular MHC scent from mice fed the same diet. Once trained, they were able to pick out that MHC scent from mice fed another type of diet. In other words, the genetically-linked BO signal persists amid stronger dietary signals and properly trained mice can find it. Chemical analysis revealed forty-nine molecules in mouse urine that vary with diet and MHC type. From these, Kwak and colleagues were able to construct a statistical model that predicts MHC type as accurately as the trained sensor mice. Pretty cool stuff.

Kwak et al. go on to claim “it should be possible to develop a detector to identify individual odortypes that can ignore environmental perturbations such as diet variation.” Once again the needle on my Bogosity Meter bounces a bit. Why? Because the mice in question weren’t recognized as individuals (Bob, Jane, John, Wendy, etc.) but as members of genetically identical groups (the Smith-family clones versus the Jones-family clones). Clonal sibship is one thing, individual identity is another. A device to “detect individual odorprints in humans” is several leaps of logic away from the results of this study.

In Fahrenheit 451, the fugitive fireman Guy Montag successfully evades the Mechanical Hound by swapping clothes with an old man, dousing himself with whiskey, and floating down a river. If pursued by trained mice from Monell, he could simply have popped a breath mint and taken it easy.

No comments: