Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Protein electrophoresis algorithm

  1. May 18, 2014 #1

    I'm trying to implement a application that will scan the electrophoresis gels and draw a graphic of the proteins: albumin, alpha1, alpha2, beta1, beta2 and gamma.
    The problem is that the resulted graphic is not as it should be. Albumin percentage is too low and the rest of the proteins are too high. Also, looking at the my graphic I can see that albumin is very low.

    I tried to make a comparative test with many applications that already exist on the web and all of them are giving me the same values as I calculated. But this values are wrong so it is clear that I'm missing something essential.

    One of this applications is ImageJ(open source).

    I have attached a scan of a normal control done on a Sebia gel.
    Correct values are:
    Albumin : 66,05%
    Alpha1 : 2,47%
    Alpha2 : 9,14%
    Beta1 : 6,9%
    Beta2 : 4,33%
    Gamma : 11,11%

    Values obtained with my app and ImageJ are:
    Albumin: 46,76%
    Alpha1 : 3,77%
    Alpha2 : 15,05%
    Beta1 : 11,78%
    Beta2 : 5,36%
    Gamma : 17,29%

    How can I reproduce the correct values?
    Thank you

    Attached Files:

    • CN.jpg
      File size:
      1.8 KB
    Last edited: May 18, 2014
  2. jcsd
  3. May 18, 2014 #2


    User Avatar
    Science Advisor
    2017 Award

    What values are you reporting? Migration distances? Calculated molecular weights? Intensities?
  4. May 18, 2014 #3
    Values are in percents and are calculated as part of the total area of the resulting graphic.
    I will edit the original post to add % sign.
  5. May 18, 2014 #4


    User Avatar
    Science Advisor
    2017 Award

    So, you are measuring the intensity of each band in the gel, then dividing by the total intensity. How do you know your correct values are correct? Are you making up a standard solution with known concentrations of each protein? Is there a way to check whether you have made this standard solution correctly?

    Another possibility is that the intensity from the albumin (the top band I presume), is saturating your detector, so your measurement is lower than the amount of protein present. Diluting the sample and running again may help (or changing the settings on your detector). How are you quantifying the intensity of each band? What type of staining are you using to visualize the proteins?
  6. May 18, 2014 #5
    The correct values are from a prepared solution, made by a specialized producer.
    Solution has these concentrations of proteins from fabrication.
    So the correct values are 100% correct.

    You are right, the scanner is not specially made for electrophoresis and it might not be sensitive enough to see the peak of the Albumin.
    I will try to fix the problem software by increasing the Albumin peak until I reach the target values.

    If Albumin intensity is the problem than I should reach all protein values at the same time as I increase albumin.

    Thank you :)
  7. May 18, 2014 #6
    So, the scanner is the problem.
    There is no complete software solution, I need a scanner that will see all the Albumin.
    Where can I find one?
  8. May 18, 2014 #7
    Have you eliminated the possibility of differential dye binding? Also if you aren't using a light box type of instrument, then you have all sorts of things to worry about. For example the lighting varying from place to place on the gel, as well as the inherent differences in the pixel sensitivities of the camera (or lens vignetting, dust etc). Some of these camera problems can be fixed easily, others not so much. You can google for something like flat-frame calibration to see what can be done with regard to the differences in the inherent pixel sensitivities.
  9. May 19, 2014 #8
    All solutions are standard, differential dye binding shouldn't be a problem but I will ask.
    Light could be a new problem. :)
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook