Wednesday, 28 January 2015

And now for something completely different

I'm scared, because my next objective is to use all the sequences I have collected so far from all three genes to make a proper tree, with a root and molecular clock. This is quite exciting, because although I have been playing around making little trees, I haven't yet done it properly. But it's also scary because what if nothing makes sense? What if the program I am using (BEAST (Bayesian Evolutionary Analysis by Sampling Trees), specifically *BEAST) doesn't work on my data? I have had that problem with countless programs and data before: it works beautifully on the example data, but not on my data. A crushing disappointment similar to when you test-drive a car and it runs beautifully, but then when you take it home it blows a head gasket. There are always solutions, but they take many long and tedious hours of removing extra spaces, inserting brackets and dots, and trying capital letters instead of lower-case letters to see if any of those work.

For the last few days I have been swept up in the completely random and unpredictable science of ecology. In particular, we (myself and my field technician) have been sampling Cantuaria habitat to see if soil moisture, soil grain size, and a bunch of other factors affect the range and positioning of populations. We've also been trying to gauge the size and range of the populations using adaptive cluster sampling (ACS), which is one of the most tedious ecological sampling techniques out there, but it's also pretty good. Basically, you decide on an area large enough to contain a few burrows (I chose a 2m X 2m square). You start off in an area with burrows, mark that there are burrows and the number of them (and other stuff you're interested in), and then move onto another area the same size but to the right of the square you've just covered and repeat the process. Then you do the same to the left of the first square, and above it and below it. If, on any of these surrounding squares, you find a burrow, you then have to sample squares above, below, and beside that square. So you can find that on two sides of one square there are no burrows, and get your hopes up and think you can now go home after friggin hours of measuring out squares and looking for burrows, but then you look in the third square and there's one burrow there so you have to look in another three bloody squares. It's a good way of finding the edges of a population, but it's also a good way of making yourself wish the populations were a little bit smaller.

Also, the day that we put out some data loggers to collect temperature and humidity data about the habitat, some cattle got into the reserve and trampled everything and destroyed burrows and pulled out probes that we'd put into burrows to see if the temperature and humidity differed in burrows as opposed to in the surrounding soil. Yet another reason that cattle should not be in NZ.

Anyway, I'd better get on and begin the process of trying to begin the process of hopefully maybe beginning to make a tree. I'll let you know how it goes.

No comments:

Post a Comment