StopTheSemanticsAnd
GetOnWithTheTestingVille
Alexandra Morton has started her dissection of the Cohen hearings and it makes for compelling reading.
One of the things, amongst many, that she wrote about today that I found interesting was the following:
“…we now have biological definitions of disease and legal definitions of disease in conflict with each other and this is not going to work…”
****
Now.
I have to be very clear here right from the outset.
I know nothing about fish biology.
But I do know a little bit about ‘expression profiling’ which allows you to identify genes that are either ‘on’ (= more expression) or turned ‘off’ (= less expression) in individual samples.
These samples could be from anything alive, be they cells in a culture dish, patients in a hospital, or animals in the wild.
Once the profiling has been done on lots of samples you can then subdivide them into groups or
clusters based their ‘on’/’off’ patterns. The fancy schmancy name for the latter is ‘gene signatures’.
Are you still with me?
Because if you are, you can pretty much understand Kristy Miller’s Science paper.
You know, the one that has caused all the kerfuffle.
****
Now, what Miller et al. did was collect tiny biopsy samples from the gills of a whole bunch of Fraser River sockeye salmon that were caught either in the ocean or in the river.
Then they did the expression profiling.
And then they did the clustering based on their gene signatures.
Then they did something very cool, which they could only do because the fish were radio-tagged before they were released after the biopsy was taken.
In short, they figured out which of the sampled fish made it to their spawning grounds and which ones did not, the latter being a measure of ‘mortality’.
And guess what….
One genomic signature stood out because it was associated with higher mortality rates, regardless where the fish were originally sampled.
Thus, it would seem that this signature, especially if it points towards something that happened to the fish way out in the ocean before they even started spawning, could be pretty important, eh?
So.
Miller et al. had look at the individual genes themselves and tried to figure out what that certain ‘something’ might have been.
And this is where things got a little speculative.
In a good way.
Because that’s the way things are supposed to work in science. First you make a series of observations and then you make a speculative prediction (ie. an hypothesis) based on those observations.
And this, from the very last line of the paper, was the Miller group’s hypothesis:
“Our hypothesis is that the genomic signal associated with elevated mortality is in response to a virus infecting fish before river entry and that persists to the spawning areas.”
Now.
With an hypothesis like that, what are scientists supposed to do?
Why, ‘test’ it of course.
And if it stands up to every rigourous test a gazillion scientists can think of, it will get to hang around for awhile.
And if it doesn’t, it will get thrown on the scrap heap where it belongs and, hopefully, all the observational data generated by all that additional testing will lead to even better and more accurate predictions/hypotheses.
And that’s how we will move forward and figure out what is really going on with the Fraser River sockeye.
All of which is just another way of saying that we will never, ever figure out what is happening by mounting PR counter-spin offensives and/or engaging in lawyer-driven semantic battles about the legaleese of disease.
So….It would appear that the only sensible thing is to do now is to unleash ALL the scientists and:
Let the testing begin!
OK?
.
