Tagged: software Toggle Comment Threads | Keyboard Shortcuts

  • kubke 15:12 on September 22, 2013 Permalink | Reply
    Tags: , , lifelong learning, software   

    Spent a large proportion of the weekend setting up the github repos. I feel like an absolute idiot at times but if I use it enough I am sure that even *I* can get a hang of it. Since I signed up for two coursera classes to learn R, I figured github would be a good place to keep the notes. One step at a time

    • kubke 21:56 on September 25, 2013 Permalink | Reply

      well – turns out that the code of conduct from coursera says I cannot share test questions and answers – so github is no a choice. Dang.

  • kubke 21:44 on November 20, 2012 Permalink | Reply
    Tags: , , , software   

    2012.11.20: Software mystery solved 

    I had this nagging feeling all evening yesterday about the software mis-reporting the injected current by a factor of 4. It seemed an odd number to built into the software – for some reason I might not be troubled if it was a factor of 10, 100 or whatever, but 4? It also seemed odd that the error would only show up in one mode but not the other…
    Anyway, I went to talk to the student to try to understand what was going on, and it turns out that at some point she needs to enter some sort of scaling value when she calibrates the system, and that is when all pieces fell into place. To enter the numbers she uses the numerical keypad at the right of the keyboard – and “4” is right on top of “1”. Well, it was the source of some laughter – and she did indeed say she had to change her scaling value from something lile 0.4 to 0.1 when she recalibrated yesterday. Problem solved. It seems it all came down to a slippy finger on the keyboard.

    But there was still the issue that her input resistances seemed a bit too high – so back to google scholar. Many papers report that they monitor the input resistance but don’t provide actual values. Then it dawned on me, as long as they show an image of a cell where they show the voltage trace we can estimate it by measuring with a ruler from the figure. And so with that new incentive she came back with a list of values from the literature. Oh my, they are all over the place, and it does not seem to matter whether they are recorded with patch or sharp electrodes. So we tried to figure out what was going on and she noticed that the time at which the steady voltage was measured varied between papers. So that might be it – is it possible they are measuring while some hyperpolarisation current is still active? So I asked her to remeasure how her estimates of Rm might vary depending on whether she measured at the end of her current pulse (as she does) or earlier in the trace. I guess I will find out tomorrow.

    In the meantime, as we were discussing all this we got into a bit of a discussion as to what was the best way to measure Rm. It became kind of funny, because we both do it slightly different and it seemed that it was the blind leading the blind for a wee while, until we realised that we were looking at the problem from a slightly different point of view, and in fact we were both right. What was good, though, is that it challenged me to think whether I could be wrong. Happy to learn I wasn’t!

    I also had the board of exmainers meeting today, but maybe I will talk about it tomorrow.
    And no, I have not done my dishes yet! Agh!

  • kubke 20:17 on November 19, 2012 Permalink | Reply
    Tags: , , , software   

    2012.11.19: One more reason not to just trust your software! 

    Another busy day – Meeting with student I am co-supervising, trying to get the HRC grant in, getting the eggs ready, and dealing with the conference organisation.

    The meeting with student went well. Turns out, she comes to me afterwards puzzled that her input resistances are different in voltage and current clamp modes, and the ones in current clamp mode turn out to look nothing realistic. I suggested that she should repeat her experiments (in both modes) with a model cell, since what she described sounded to me like the software was off by an order or magnitude or something – though it was weird that the problem would only be there in one of the modes. Anyway, she did this and was skipping with joy when realising pClamp was off by a factor of 4 (whatever she meant by that).

    Sort of timely given my recent blog post (http://blogs.plos.org/mindthebrain/2012/11/15/failure-to-replicate-as-an-opportunity-from-learning-2). And underscores what I always say: do not trust software, not even commercial software! test test test and make sure software does what it says it does. When I think about it, time to look to see if anyone else has noticed this and whether some of the publications using pclamp are off!!!! Oh my…..

    Glad that puzzle has been sorted.

    Anyway, eggs are set, and should start fixing embryos on Friday – must remember to finish doing the dishes

Compose new post
Next post/Next comment
Previous post/Previous comment
Show/Hide comments
Go to top
Go to login
Show/Hide help
shift + esc