Yes, you need to study a bit to do it right, and even then, there are some sources of systematic error that are difficult to avoid. In my opinion, SDS-PAGE/densitometry cannot be used as anything more than a rough guide to make comparisons between a limited number of experimental conditions. Some of the things you have to account for:

- When imaging, you have to ensure that the picture is not saturated. This should be taken care of by your photographic or densitometric setup.

- Non-linearity of the stain/Narrow dynamic range. This is an important problem for Coomassie, not so much when using fluorescent stains.

- Disparities in signal strength for different proteins. Again, an important consideration when using Coomassie.

The problems posed by a narrow dynamic range cannot be overstated. Oftentimes you are dealing with intermediate purificates, where the difference in concentration between the target protein and contaminants may be large. In that case, loading a lot of protein to accurately quantify the contaminants pushes the signal from the target protein into saturation, and if you try not to overload the gel with your target protein, then you underestimate the amount of contaminants due to lack of sensitivity.

I don't know your requirements regarding time/money to be spent, but if you want accurate/reliable results, the best solution is to use an ELISA for the target protein and a separate ELISA for total host cell protein.

The gel:

Hi Peter

You said that your protein is expressed in samples loaded on lane 6-9. However, the mol wt of the bands are different from the standards.

Anyway, in order to quantify the bands you need to get histogram of bands from each lane separately. I guess you are using ImageJ software. The curve would be bell shaped. The area under the curve would represent the band density.

Hope it helps!!!

Thanks for the help, but I am bothered more with the exact details. I found out that it matters which image file format one is using (JPEG loses information, for example). Densitometry is not as easy as it seemed to me on the first glance, I suppose I need to delve into image processing to do it right.

http://imagej.1557.x6.nabble.com/Incorrect-density-values-when-quantifying-PNG-files-exported-from-GIMP-td5000862.html

Yes, you need to study a bit to do it right, and even then, there are some sources of systematic error that are difficult to avoid. In my opinion, SDS-PAGE/densitometry cannot be used as anything more than a rough guide to make comparisons between a limited number of experimental conditions. Some of the things you have to account for:

- When imaging, you have to ensure that the picture is not saturated. This should be taken care of by your photographic or densitometric setup.

- Non-linearity of the stain/Narrow dynamic range. This is an important problem for Coomassie, not so much when using fluorescent stains.

- Disparities in signal strength for different proteins. Again, an important consideration when using Coomassie.

The problems posed by a narrow dynamic range cannot be overstated. Oftentimes you are dealing with intermediate purificates, where the difference in concentration between the target protein and contaminants may be large. In that case, loading a lot of protein to accurately quantify the contaminants pushes the signal from the target protein into saturation, and if you try not to overload the gel with your target protein, then you underestimate the amount of contaminants due to lack of sensitivity.

I don't know your requirements regarding time/money to be spent, but if you want accurate/reliable results, the best solution is to use an ELISA for the target protein and a separate ELISA for total host cell protein.

Dear all,

thank you again for all your input! Meanwhile I have prepared a publication covering this interesting topic. I would be glad if it could be of any help, so I wanted to point to it here:

https://www.researchgate.net/publication/269701198_Integrated_protocol_for_reliable_and_fast_quantification_and_documentation_of_electrophoresis_gels

Please also recommend this question/answer and the corresponding publication, so that others may find it more easily.

If you find it useful for your work, please cite the paper in your publications and share. Comments on the software are always welcome, either here or via private message.

Best,

Peter

Article Integrated protocol for reliable and fast quantification and...

I'm interested in this as well. What's the best method to take the "photo" of the gel? Using a normal scanner is a good option? Thanks.

Hi David,

you can use a sophisticated Gel documentation setup if available at your Institution, but a common flatbed scanner is totally fine. Just make sure the Gel receives a constant light intensity over its complete area and that the Image resolution delivered by the Scanner is high enough. Also make sure to use TIF.-format for scanning if possible (uncompressed data for archiving) and convert to PNG for quantification.

The cheap and very compact scanner I described in the publication is meanwhile unavailable on the market (at least in Germany) as the manufacturer realized that people are using the device rather for scanning gels than photos and is now selling the same scanner as "Gel documentation system" but for a price two magnitudes higher... That's marketing...

Best,

Peter

Hi Peter, thanks for your answer. I am going to read your paper as well. I'm going to take into account your information about the parameters to be careful about. 

Alok Patel thanks! I've been using ImageJ, but not like this way. Gonna check it out deeply.