2/11 Lines of sight

In which we learn how Hoagland's rules, lines down the middle of the root may wobble, and pinholes might look sharp but reveal little...

Last week is even more of a blur than usual because my partner and I went to London. Even though we were there for just Friday and Saturday, it seems like a week, what with everything we saw, smelled, tasted, and heard. One of those sights was a poster for the Great British Bioscience Festival in a tunnel under St Pancras Station, notable because we’ll be there at CPIB’s booth trying to explain why it is pretty cool to give a root a CAT scan.

Now to the week that was. I gave a departmental seminar on the Wednesday, which preoccupied me, as well it should.

The only wet lab work that I did was to tick the position of the roots on the plates, comparing the standard issue one-half strength “MS” medium to Hoagland’s (I described the rationale for this experiment on 19/10). While it is always dangerous to forecast without quantification, what I saw by eye is enough to justify a comment. Or two. I had put a pair of plates in two spaces: “Darren’s” growth chamber (A45 in the graph here), and “Mandy’s”, A61, which I hadn’t tried in the last experiment and as its number suggests, is right next to A60. In A35, the roots on the Hoagland’s grew a little faster than the ones on MS, which feeds into my suspicions about the default medium. But in Mandy’s chamber, while the roots on the MS plates were grinding out a few mm a day, looking a lot like what I saw previously for A60, the roots on Hoagland’s had found their inner nuclear reactor and were flying, at least a centimeter per day, maybe more. This is about as fast as I have ever seen roots grow.

Well good! My goal of having roots grow rapidly for making movies looks like it can be met, provided this wasn’t some cosmic ray-induced fluke. I have been reading about solar flares and giant sunspots. I plated a couple of plates and I will see about the reproducibility. Another question is whether the crawling growth on MS is for real. Maybe I made a crap batch of plates? Given it was the first batch of plates I have made in about a century, why not? Because I have opened my mouth to everyone about the growth rate problems (sharing the graph for example), I should find out. After crunching the numbers, Monday or Tuesday, I will ask for volunteers to provide an MS plate so I can compare plates made by others to mine and see if the poor growth is due to icky old MS or to my ineptitude.

I tracked down a glass still and got a liter of still water because in the past when I experienced slow root growth, switching to a still was the magic bullet. Based on how well roots seem to be growing in Mandy’s on Hoagland’s, I will put off trying that wrinkle, at least for now.

In other news, I have finally had a chance to crunch the movies I took of botero roots. As described earlier (12/10), botero is a mutant suspected of impaired mechanical connections between cells, and insofar as those connections represent a way in which the velocity profile could be stabilized, it seemed reasonable to make movies of this mutant. So far, the profiles seem every bit as rock solid as those of the wild type, but I have only looked at three. This makes the relevance of mechanical connections look doubtful, at least the ones disrupted by the mutant. But I contacted a botero connoisseur (of the mutant, not of the artist) who mentioned that the line I got from the stock center harbors a point mutation in the gene, which might allow the line to retain some botero activity. He promised to send me seed of a line that has a whacking great transposon sitting in the gene, one that he has verified wipes out all traces of the gene’s RNA. Since I need a larger sample size anyway, making movies of this so-called null allele will be useful.

I might have crunched the fourth botero movie but I got distracted by doing some tests to improve how the software I am using calculates the velocity profile. I can measure the stability of the velocity profile only to within the noise limit set by the software. As it goes now, the user has to input a few points that define the midline of the root. I have noticed that choosing different points gives rise to different velocity profiles. The differences are small, and for previous tasks not likely to cause any hassle, but in my attempt to crank accuracy to the max, these discrepancies are annoying. I looked at what happened by choosing points in different ways and I discovered a problem with how the program is drawing the midline. I am going to work with Gang Dong who wrote the software to try to improve it.

The stem-segment growth set up also claimed some of my attention last week, but mostly in a paper-and-pencil kind of way. As I explained last week (26/10), I felt as though I had arrived at a satisfactory arrangement of light, sample, and camera, and I had asked Darren to do the final step of building a holder that will allow me to place the washer containing the stem segments in front of the camera, reliably, in the dark. But there was one thing that worried me. Maybe not enough to keep me from sleeping but a worry all the same. To achieve the focus shown in the image, I had to close the iris of the lens all the way down. In microscopy when you do that, the quality of the image (resolution) goes to pot. Maybe that was happening here?

I contacted a friend of mine who is an optics expert and he gave me a way to look at the optical resolution of the system compared to the number of pixels in the camera. Following his formula, it turns out the set up will give me about two pixels per optically resolved unit, which is borderline acceptable. I’d like more. Without getting into the equations, it boils down to the relationship between the diameter of the iris and the distance between the lens and sample. Opening the iris will mean that the sample will have to be moved farther from the lens. Opening the iris gains resolution, moving the sample reduces it. But symmetrically? If doubling the iris diameter forces me to double the distance between lens and sample then it is a wash. But if I can open the iris more than moving the sample, then I gain resolution. In need to try this before getting locked into a fixed configuration. So Darren is off the hook until I can sort this out. Rolling up sleeves on Monday.