Print

Replicate-Based Noise Corrected Correlation (RBNCC)

J. Adler, S.N. Pagakis & I.Parmryd - Uppsala University

Introduction

RBNCC is an option for the Colocalization Analyzer to correct for noise when calculating the Pearson and Spearman colocalization coefficients between two channels of an image. Noise is unavoidable and introduces a broad range of uncertainty leading to underestimation of colocalization. Noise correction by RBNCC is made from two images of the same specimen. Colocalization coefficients are calculated and directly corrected with RBNCC when this option is activated. (The RBNCC method has been developed by J. Adler, S.N. Pagakis and I. Parmryd from Uppsala University; Journal of Microscopy 230, 121-133; 2008).

Colocalization Correction

The correction is based on calculating the influence of noise of an image. To do this one needs to acquire two images of a specimen, such that the object are still the same, but the noise is different. Depending on which coefficient you would like to correct for, the colocalization between the corresponding channels are evaluated. The diagram below shows the computation process.




The two sets of two-channel images allow for four measurements of the correlation (Pearson or Spearman) between the channels. The individual measurements and their mean (Rmgr) are however affected by noise. In the absence of all forms of noise a replicate image would be identical to the original image.

The quality measures are the colocalization coefficients (either Pearson of Spearman) between the red channels (Rrr) and the green channels (Rgg) and are combined into a joint correction factor (Cgr).
C_{gr} = \frac {1} {\sqrt{ R_{rr} \cdot R_{gg}}}


This correction factor is then applied to the mean correlation between the channels (Rmgr) and produces a correlation measurement free from the influence of noise (Rc).
R_{c} = R_{mgr} \cdot C_{gr}


Colocalization Analyzer

Image The RBNCC is implemented in de Colocalization Analyzer in the Huygens Software. One can turn on the RBNCC calculations and select the second image. By simply pressing Compute the coefficients are calculated and directly corrected with RBNCC. The colocalization correction is reported behind Colocalization correction:
  • P shows the Pearsons correction
  • S shows the Spearman correction
  • OP shows the Object Pearson correction
  • OS shows the Object Spearman correction


Check for more info