I've been blotting samples derived from cells that were treated with a ligand. I expect this ligand to activate a signaling pathway, and I expect certain components of this pathway to be phosphorylated when active. To assay for activation of those components, I ran gels loaded with two lanes of each sample. I then cut the blots in half so that I had two identical blots. One was immunolabeled with antibodies against the total protein of interest, and one was immunolabeled with antibodies specific to the phosphoprotein of interest. They were detected and imaged side by side. I'm trying to extract quantitative data from these.
Alas, sometimes - and I don't understand why - the phosphoprotein blot has bands that are darker than the total protein blot. Should these blots necessarily be excluded from my analysis? If not, how should I go about the image analysis, and how can I normalize the data from that blot to the rest of my quantified blots such that I can generate a standard deviation that isn't absurd?
So, usually you must repeat the experiment several times. If you get the same result, then you cannot exclude from analysis. If you do not get the same result then it could be easily a loading error (loading too much protein by accident).
Also sometimes your exposure could be too much (saturated). We use horseradish peroxidase and we add substrate. If it is too much it will appear too dark. Using a milder substrate may help.
The ponceau stains I've done seem to rule out loading issues. I've repeated several times, and the issue with the phosphoprotein bands being darker seems sporadic - sometimes they're darker, usually they're not. It was this observation that lead me to think that perhaps this has little to do with affinity for the antigen. I was hoping to avoid reprobing all these blots for a housekeeping protein, but it sounds like that may be my only option. I appreciate your feedback.
I'm particularly interested in if it is possible to transform the data I've gotten from these blots. For example, for my control sample I have ratios of phosphoprotein/total protein of 0.68, 8.60, 0.42, 0.56, and 1.25. The pattern that I see in my treated samples relative to this control is the same from blot to blot (both my samples similar to control, looks like there's no effect), but the ratio of phosphoprotein/total protein is all over the place from blot to blot.
All of the above responses are consistent - I would also strip and probe both lanes (phosph and non-phos) with a housekeeping protein and assess whether "normalizing" to a housekeeping protein gives less "shot-gun" results. Good luck.
- Badges
- Science method