Thanks for your reply.
Of course, I should have mentioned that I do not want to determine
absolute protein quantities but only relative quantities for each gel,
by scaling the maximal value on each gel to one, for instance. Just to
get some numbers.
Moreover, I assume that I use the same loading for each lane and that
the response is not saturated.
When I understood you correctly you propose the following:
1) Quantify the band for a specific lane (e.g. by counting black pixels
in a rectangle around the band).
2) define the background for this lane (e.g. by counting black pixels in
a rectangle placed next to the one above but in the same lane)
3) Substract the the two values.
4) do this for each lane.
is that right?
thanks for your reply,
joerg
DK schrieb:
> In article <4fd7h3F1i4v77U1 at news.dfncis.de>, Joerg Schaber <schaber at molgen.mpg.de> wrote:
>>>Hi,
>>>>I am not an molecular biologist but use Western Blot data for modelling.
>>I quantify gels using MultiGauge to get numbers for the band intensity
>>on the gels.
>>>>My question is whether there exist some standard procedure or 'good
>>scientific practice' concerning quantification of Western Blot data.
>>>>For example what my molecular biology colleagues told me was to define a
>>background to each band on the gel. But then, to get the quantification,
>>some just subtract the value for the background from the value for the
>>actual band, whereas others first scale the value of the band by a
>>factor calculated from the mean background before substracting the
>>background. These different scaling methods can yield quite different
>>results.
>>>>I would be grateful for some hints or references or maybe that's not the
>>right newsgroup at all?
>>> And you do not even mention the most important thing.
>> Band intensity on a membrane is *not* a linear function of the protein
> amount! With *all* detection techniques the response saturates. The
> precise function is not known in advance and its parameters vary
> considertably experiment to experiment. Therefore, if one wants to
> make quantitative statements from Westerns, there *must* be
> a standard calibration curve that covers the entire range of your
> experimental response. This is not done in 99.9% cases because, in
> reality, most people who do Westerns can't care less about quantitative
> accuracy of their results. Without calibration, however, any
> fancy-shmancy software produces only an equivalent of "there
> is more here than there".
>> As far as background problem is concerned: The approaches I've seen
> in different programs vary significantly and in the end the differences
> between them don't matter too much because the accuracy of gel
> loading/sample handling is almost invariably worse than the errors in
> background determination. It matters even less when you have a
> calibration curve determined using the same method.
>> IMHO, the correct way of handling background is to treat it as a
> local value. Using means for two lanes that have different loading
> per lane makes absolutely no sense to me. The main source of
> the background is proteins spreading/smearing in the gel lane,
> not non-specific response from areas totally devoid of proteins.
>> DK
>>>>
