How to interpret data generated from Intensity Correlation Analysis?
Locked
How to interpret data generated from Intensity Correlation Analysis?
 
|
I am new to confocal analysis (Infact I have observed most of the people,
especially graduate students shying from quantification saying that there is no tool to quantify the images). I want to make a quantitative meaning of my co localization studies which I am doing in primary hippocampal neurons. The important thing is that I know there is a very small percent of my two proteins co-localizing, but how to make it understandable in terms of words. I performed Intensity correlation analysis, and obtained following data for my protein sets. Truly speaking, looking at values, I could not interpret anything. I therefore request the experts to help me get the meaning from my sample data. Pearson´s correlation coefficient Mander Overlap Coefficient Green:Red Ch1:Ch2 Mander´s colocalization coefficient for Ch1, M1 Mander´s colocalization coefficient for Ch1, M2 No. Of +ve pixels No. of pixel pairs (atleast one above zero), Ntotal Intensity Correlation Coefficient, ICQ Threshold of Ch1 Threshold of Ch2 Ref 1 0.367 0.944 0.685 0.518 0.338 1186 18689 0.131 59; 243 38; 251 Ref 2 0.025 0.877 0.454 0.601 0.293 2560 45253 0.041 31; 246 26; 252 Ref 3 0.064 0.897 0.572 0.581 0.371 1671 36902 0.009 52; 248 41; 252 Ref 4 -0.02 0.865 0.588 0.509 0.309 1607 38968 0.02 36; 248 39; 249 Also What does Ntotal means in the plugin? Regards Amulya Nidhi Shrivastava, B.Engg. Graduate Student Sieghart´s Group Center for Brain Research Medical University of Vienna, Austria Ph: +43 1 4277 62953 Mob: +43 6505 420 301 Fax: +43 1 4277 62899 Web: http://phd-cchd.at |
Re: How to interpret data generated from Intensity Correlation Analysis?
|
|
Hi Amulya,
To start you will need to go to the original papers to understand how the values are calculated and how they can be interpreted. Li Q, Lau A, Morris TJ, Guo L, Fordyce CB, and Stanley EF (2004) A Syntaxin 1, G{alpha}o, and N-Type Calcium Channel Complex at a Presynaptic Nerve Terminal: Analysis by Quantitative Immunocolocalization. J Neurosci, 24, 4070-4081. Manders EM, Stap J, Brakenhoff GJ, van Driel R, and Aten JA (1992) Dynamics of three-dimensional replication patterns during the S-phase, analysed by double labelling of DNA and confocal microscopy. J Cell Sci, 103, 857-862. Manders E, Verbeek FJ, and Aten JA (1993) Measurement of co-localisation of objects in dual-colour confocal images. Journal of Microscopy, 169, 375-382. Tony Tony J. Collins, Ph.D. McMaster Biophotonics Facility Dept. Biochemistry and Biomedical Sciences HSC 4H21A McMaster University, Hamilton, ON, L8N 3Z5 [hidden email] www.macbiophotonics.ca > -----Original Message----- > From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of > Amulya Nidhi Shrivastava > Sent: May 6, 2008 4:42 AM > To: [hidden email] > Subject: How to interpret data generated from Intensity Correlation > Analysis? > > I am new to confocal analysis (Infact I have observed most of the > people, > especially graduate students shying from quantification saying that > there is > no tool to quantify the images). I want to make a quantitative meaning > of my > co localization studies which I am doing in primary hippocampal > neurons. The > important thing is that I know there is a very small percent of my two > proteins co-localizing, but how to make it understandable in terms of > words. > I performed Intensity correlation analysis, and obtained following data > for > my protein sets. Truly speaking, looking at values, I could not > interpret > anything. I therefore request the experts to help me get the meaning > from my > sample data. > > > > > > > Pearson´s correlation coefficient > > Mander Overlap Coefficient > > Green:Red Ch1:Ch2 > > Mander´s colocalization coefficient for Ch1, M1 > > Mander´s colocalization coefficient for Ch1, M2 > > No. Of +ve pixels > > No. of pixel pairs (atleast one above zero), Ntotal > > Intensity Correlation Coefficient, ICQ > > Threshold of Ch1 > > Threshold of Ch2 > > > Ref 1 > > 0.367 > > 0.944 > > 0.685 > > 0.518 > > 0.338 > > 1186 > > 18689 > > 0.131 > > 59; 243 > > 38; 251 > > > Ref 2 > > 0.025 > > 0.877 > > 0.454 > > 0.601 > > 0.293 > > 2560 > > 45253 > > 0.041 > > 31; 246 > > 26; 252 > > > Ref 3 > > 0.064 > > 0.897 > > 0.572 > > 0.581 > > 0.371 > > 1671 > > 36902 > > 0.009 > > 52; 248 > > 41; 252 > > > Ref 4 > > -0.02 > > 0.865 > > 0.588 > > 0.509 > > 0.309 > > 1607 > > 38968 > > 0.02 > > 36; 248 > > 39; 249 > > > > Also What does Ntotal means in the plugin? > > > > Regards > > > > Amulya Nidhi Shrivastava, B.Engg. > > Graduate Student > > Sieghart´s Group > > Center for Brain Research > > Medical University of Vienna, Austria > > Ph: +43 1 4277 62953 > > Mob: +43 6505 420 301 > > Fax: +43 1 4277 62899 > > Web: http://phd-cchd.at > |
|
|
In reply to this post by Amulya Nidhi Shrivastava
Hi Amulya,
I'm glad to see you found and tried the ICA analysis. I'm the creator of this analysis method and Tony (who replied above) did most of the work creating the plugin for ImageJ making it easy to carry out. Of course, one still has to interpret the data. Tony is right in that if you read the original paper by Li et al in J. Neuroscience the principle behind ICA analysis is explained. Tony also added older analysis methods to the plugin, making it quite a comprehensive analysis approach - but also perhaps a bit intimidating. The difference between the older analyses methods and the ICA is that the former either rely on thresholding (deciding if a pixel is 'stained' (positive) or 'not stained' (negative) rather than using its actual intensity value and they do not truly test whether the two staining intensities vary in synchrony or not. The ICA calculation gives you input not only on whether the two stains are in the same place (which is rather subjective since you have to decide on a threshold) but if they vary together - as they should if the two proteins are parts of the same complex. The diagrams can also tell you if the two proteins co-vary in one region of the cell while not doing so in another. Note that ICA analysis is only useful if you look at a single structure - you can not do it on multiple cells at the same time as you will then have regions with cytoplasm and regions without - and that comparision will almost certainly give you a strongly positive value. So the first thing is to identify a region of interest - the part of the cell that you want to test for covariance. Once you have done the analysis look for the ICQ value. This is a single value 'average' of the covariance. In the series of analyses you present only the first one is strongly positive. On average the data suggests that your two stains (proteins) do not covary in the region of interest that you were looking at. You can go back to the cells and look at a more restictive region if you like - non-nucleus; within the golgi etc if you can identify these regions with some comfidence. Please let me know if you have any more problems/questions. Elise
|
«
Return to ImageJ
|
1 view|%1 views
| Free forum by Nabble | Edit this page |