Since you are writing about calibration data, I assume you are running
samples of control DNA with known concentration on the same gel as your
unknown samples, and you make sure that EtBr stains most of your DNA.
I myself would find it more convenient (and correct!) to use a "Measure"
macro, in particular its "Measure Sum of Pixel Values" function.
Most correct results will be obtained when your adjust the camera and
brightness/contrast settings so that the background is ~ zero pixels (or 255
in case of invert picture, since pixels can have values no more than 256 =
2^8 bits). After grabbing the picture into a file, use freehand or
"semi"-freehand select tool to outline the spot. Select the macro function
from the Special menu ("Measure Sum of Pixel Values"), press Ctrl-1 (= Do
Measure) and read the calculated values of mean pixel value and number of
pixels. One magnified by another gives you sum of all pixels (mean = sum of
pixels / number of pixels), which is equivalent to the volume of the peak =
quantity of DNA in the spot. As you understand, it is not necessary to be
very careful in outlining the band - take as much the area adjacent to the
spot as reasonable - it would not contribute to the final value of the sum
of pixels as soon as the value of background pixels ~ 0. If it is very
difficult to adjust settings so that the background is 0, subtract the
measures mean value of the background from the whole scan before doing
Hope this helps.
"mcalexp" <mcalexp at my-deja.com> wrote in message
news:947uq9$4q0$1 at nnrp1.deja.com...
> Hello everyone,
>> I am new to using NIH Image so this might
> be a silly question. I want to use Image to
> quantify the amount of DNA in some samples. I
> fit the calibration data to a Rodbard-type curve
> and then use the gelplot macro to mark the lanes
> and obtain areas for the bands of interest. Is
> there a way for NIH Image to convert these areas
> to the units used for calibration (e.g., ng DNA)
> or are the gelplot macros used only for
> determining ratios? Is there a macro that would
> let you calibrate using the areas (instead of the
> 8-bit gray-scale values) under the peaks of the
> bands used for calibration? Is it accurate enough
> to use the gray-scale value (instead of the area)
> of the band in question along with the calibration
> curve to determine the amount of DNA (or whatever
> else one is trying to quantify)?
>> Any help would be appreciated. Thank you
> all for your time.
>>>> Sent via Deja.com