;)
Q: How does REDFIN normalize DIGE projects?
Posted by ludesi in 2DE Knowledge Base
Ever wanted to know the nitty gritty of what happens behind the scenes in REDFIN?
Well, if the DIGE normalization procedure in REDFIN has been one of those things you’ve always wanted to know more about, then I dare say you’re in for a treat…
Ok, here goes.
Gaia: “Those rats aren’t normal. I suspect foul play.”
Captain Planet and the Planeteers” (1990)
Normalization is the process of making these spot volumes comparable between gel images in the face of technical differences in staining, scanning, sample volume, and so on.
In projects that use a DIGE methodology, REDFIN applies a two-step normalization procedure.
Step 1 normalizes all spots within a DIGE gel using a method of normalization that works with gel pairs. The algorithm computes a normalization factor for each of the images obtained from that particular gel and uses it to normalize the spot volumes across all three images:
• For each image pair that has been matched, all matches between the two images are extracted. The top 50% matches, based on the spot volumes, are kept.
• A spot volume ratio is calculated between each kept spot pair match. A pair-wise normalization factor between the images is computed by taking the median of these ratios. This ensures robustness and accuracy even if the project contains up to 50% regulated spot volumes.
• Each image normalization factor is computed by optimizing an over-determined equation system, resulting from the pair-wise normalization factors.
• Finally, the spot volumes in each image are multiplied with the corresponding image normalization factor.
Step 2 normalizes all spots between the DIGE gels.
• An average internal standard is created using the normalized spot volumes (from step 1) of all internal standards.
• For each protein in each gel a normalization factor is calculated as: spot volume in the average internal standard divided by the corresponding spot volume of the internal standard in that gel.
• Every spot volume in every image is then multiplied by its normalization factor.
Professor Sharp: “Wanna bet? Scarab used a simple matter transducer for the transformation. It’ll be easy to reverse. All we have to do is have a fully charged, sonic-specific bionic being commit tactile energy transference.”
“Bionic Six” (1987)
In a data transformation step, the spot volumes are standardized by calculating the ratio between the spot volume in each image and the volume in the internal standard. These standardized values are log-transformed, and the transformed values are used in subsequent statistical tests.
However, for direct volume comparisons or Fold Change calculations, the original (normalized) spot volumes from each image are still used.
Everything crystal clear now? Awesome.
;)
