Systematic errors could be identified by correlating results with other DNA sequencing results.

Using a large sample, like the proposed Personal Genome Project [unsure if they have gotten in touch with any of those who expressed interest in participating] could be useful in showing any systematic mis-reads, as long as the Personal Genome Project is using another method to sequence the participant's DNA.

I have narcolepsy, which might preclude me from any normal research into sleep-Internet use.

I write email and forum posts in my sleep on a semi-recurring basis [the more stressed out I am and physically exhausted, the more often it occurs].

It is really interesting [and kind of terrifying, the "oh dear, what did I write this time, and who to?" sinking feeling of realization when I wake up to different browser page open than expected] to see the fluidity of the words I write. More free-flowing, less choppy, but still recognizably my style of writing and thoughts.

I think it would be really interesting if there was a way that an ongoing polysomnography could be preformed, to correlate with what I post/write and when. The sleep lab that my doctor uses does not allow computers to be on when testing, lights out at 10-11pm, and up at 7am -- which does not at all represent my typical sleep behavior of falling asleep somewhere around 2am, getting up between 10-12noon, generally curled around laptop with password-requiring screensaver.

Passwords really aren't that hard to type in sleep - it is patterns that you memorize of how your hands move, when you type them in enough. The interesting part is that she had her hands aligned on keyboard properly to be hitting correct keys in sleep -- some of the stuff I've typed in sleep looks like gibberish, as the nubs on my F and J keys are worn down, so I guess I've ended up with hands slightly over from where they should be.

This got my attention and curiosity, especially since an initial glance at google didn't seem to yield an easy answer.

I ended up finding Secure Continuous Remote Alcohol Monitoring - its a anklet that measures the amount of alcohol in perspiration, intended for use by court system with DUIs and such.

From the example charts here and here, it looks like alcohol is detectable in perspiration for around 3-4 hours after the amount of alcohol in perspiration stops rising, which is likely slightly offset from the time at which the person stops drinking. I suspect there is a delay between alcohol in mouth to exiting the body via skin - the site does not seem to state the amount of time between a monitored individual imbibing alcohol to its detection in perspiration.

Axe does contain a bit of alcohol, possibly this is continuing to trigger her sense of smell of 'booze' - particularly if you and she had ever encountered each other at a point in the past where you may have smelt of both Axe and booze from previous night, or happened to wear the same socks you were wearing while drinking.

Hrm-- I wonder if alcohol being perspired through feet would transfer through socks and get ingrained in shoes.