What a good idea! It's been years since I did any Western blotting and at the time we analysed the blots with a Densitometer. Will be interesting to check out more current methods and if your technique here is widely used among blotters, or could add to the range of tools, or even solve specific problems.
Thanks Dorothy. I am genuinely interested in how people do this. I’ve seen many a blot over the years but never had to measure any myself. It always seemed a bit ‘iffy’ in terms of exactly what and where to measure, particularly when lanes can overlap. Hopefully we can get a discussion going here. Good luck with your own channel. ‘Step into Science’ is a great title.
A while now since I was into this. I think quantifying gels is very difficult and that a failure to really appreciate this difficulty is one of the major problems that has made many studies based purely on blotting very difficult to reproduce. I think an important thing is to load an extra lane which consists of equal amounts of every sample mixed together. This lane, by definition gives you an average signal....since it consists of equal amounts of every individual protein sample on the gel. This gives you an internal standard that you can relate every sample to. Virtually nobody does this.... Normalizing to the abundance of a cellular protein such as actin is often done. However...I would say that over half of the gals shown in many papers show actin bands that are solid black.....completely over exposed. This means that they completely fail to do what they set out to do. Anyway....great video but I'm retired now!!
Hi Stuart, thanks for chipping in. I agree about the exposure. We go to great lengths in microscopy and image analysis to make sure our images fall within the 8-bit range but avoiding saturation. I’ve assumed that you can really only measure size of a blob and that intensity is probably meaningless unless you have a control as you suggest. Retirement sounds nice. Cheers. C.
I asked this questions just a few months ago.I've posted a tutorial of western blot quantification using the 'integrated density method' in Analyse. However, I noticed that for faint bands, it was not very accurate. I then used the area under the peak method (recommended in the NIH imageJ docs) and it was more accurate.
Hi Adwoa, looks like we had the same idea around the same time. It’s an interesting topic. On very bright/dark bands I suspect pixel values will be 255 or 0. In that case measuring intensity doesn’t tell us much and you could just count number of pixels above a background value. Faint bands are more interesting in trying to decide on the best technique. Good to have the discussion though. I liked your video. C.
Ooh..controversial topic! I always used the threshold technique, because like you I did a lot of image analysis, and this method made sense to me. I used it only when I wanted to show the fold difference of expression of a protein between samples, when visually you could already see quite a difference. It helped back up the data when you could show similar values between repeat blots from different experiments. It’s like Marmite though, some people love putting values to westerns, others hate it!
That’s a good way to look at; only measure in specific circumstances. From what I can tell in the literature, it looks like each experiment would only have a few blots. We’re not talking about having 100 blots to analyse. In that case, manual thresholding and the visual cortex are probably the way to go. Thanks.
I don't understand why you change the threshold. It seems that should be constant. BTW, your bands are overexposed, so you won't be able to properly quantify amount of material in your bands.
Hi, I agree on the over exposure. I don’t really do this type of work and these blots (from an undergraduate student project) were all I had access to. Only using them to work through the types of analysis. Thanks for watching though. C.
Hi, You would use the 'Edit/Selection/Specify' option to make the first box. Then go to 'Analyse/tools/ROI manager'. Select 'ADD' and it will show your specified Region Of Interest. From there you can then apply that exact region to other locations or other images. C
What a good idea! It's been years since I did any Western blotting and at the time we analysed the blots with a Densitometer. Will be interesting to check out more current methods and if your technique here is widely used among blotters, or could add to the range of tools, or even solve specific problems.
Thanks Dorothy. I am genuinely interested in how people do this. I’ve seen many a blot over the years but never had to measure any myself. It always seemed a bit ‘iffy’ in terms of exactly what and where to measure, particularly when lanes can overlap. Hopefully we can get a discussion going here. Good luck with your own channel. ‘Step into Science’ is a great title.
A while now since I was into this. I think quantifying gels is very difficult and that a failure to really appreciate this difficulty is one of the major problems that has made many studies based purely on blotting very difficult to reproduce. I think an important thing is to load an extra lane which consists of equal amounts of every sample mixed together. This lane, by definition gives you an average signal....since it consists of equal amounts of every individual protein sample on the gel. This gives you an internal standard that you can relate every sample to. Virtually nobody does this.... Normalizing to the abundance of a cellular protein such as actin is often done. However...I would say that over half of the gals shown in many papers show actin bands that are solid black.....completely over exposed. This means that they completely fail to do what they set out to do. Anyway....great video but I'm retired now!!
Hi Stuart, thanks for chipping in. I agree about the exposure. We go to great lengths in microscopy and image analysis to make sure our images fall within the 8-bit range but avoiding saturation. I’ve assumed that you can really only measure size of a blob and that intensity is probably meaningless unless you have a control as you suggest. Retirement sounds nice. Cheers. C.
I asked this questions just a few months ago.I've posted a tutorial of western blot quantification using the 'integrated density method' in Analyse. However, I noticed that for faint bands, it was not very accurate. I then used the area under the peak method (recommended in the NIH imageJ docs) and it was more accurate.
Hi Adwoa, looks like we had the same idea around the same time. It’s an interesting topic. On very bright/dark bands I suspect pixel values will be 255 or 0. In that case measuring intensity doesn’t tell us much and you could just count number of pixels above a background value. Faint bands are more interesting in trying to decide on the best technique. Good to have the discussion though. I liked your video. C.
Ooh..controversial topic! I always used the threshold technique, because like you I did a lot of image analysis, and this method made sense to me. I used it only when I wanted to show the fold difference of expression of a protein between samples, when visually you could already see quite a difference. It helped back up the data when you could show similar values between repeat blots from different experiments. It’s like Marmite though, some people love putting values to westerns, others hate it!
That’s a good way to look at; only measure in specific circumstances. From what I can tell in the literature, it looks like each experiment would only have a few blots. We’re not talking about having 100 blots to analyse. In that case, manual thresholding and the visual cortex are probably the way to go. Thanks.
I don't understand why you change the threshold. It seems that should be constant. BTW, your bands are overexposed, so you won't be able to properly quantify amount of material in your bands.
Hi, I agree on the over exposure. I don’t really do this type of work and these blots (from an undergraduate student project) were all I had access to. Only using them to work through the types of analysis. Thanks for watching though. C.
Thanks a lot!
How do you duplicate the rectangles so that they're the same size?
Hi, You would use the 'Edit/Selection/Specify' option to make the first box. Then go to 'Analyse/tools/ROI manager'. Select 'ADD' and it will show your specified Region Of Interest. From there you can then apply that exact region to other locations or other images. C
After making a selection and taking a measurement, you can use the arrow keys to move the rectangle to a new location, measure, repeat.