Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Correlation statistics for cDNA microarray image analysis
Nagarajan R., Upreti M. IEEE/ACM Transactions on Computational Biology and Bioinformatics3 (3):232-238,2006.Type:Article
Date Reviewed: Jun 18 2007

This paper proposes the use of correlation statistics to analyze spots in cDNA microarray image analysis, and the authors use these techniques to segment spots in the images (separating the spot’s foreground from the background). First, the authors investigate the use of Morgera’s covariance to analyze the nature of the correlation into the microarray spots, so they can quantify the correlations of the spots’ pixels in Cy3 and Cy5 channels (for genes that are differentially expressed in one of these channels and for genes that are not differentially expressed in these channels, because they have an equal abundance--equal or similar spots in both channels--or because they are not present at all--no spots in both channels). Then, the authors propose the use of two correlation techniques, Pearson correlation and Spearman rank correlation, to segment the microarray images, separating the foreground pixels from the background. The authors compare their segmentation method with other well-known methods found in the literature of this field: PAM, k-means, and SPOT (region growing). The results, according to the authors, indicate that their proposed method is able to reduce the number of false-positives (incorrect flagged spots identified in the segmentation process).

Although the paper is well written, it is somewhat short, does not describe very well the real contribution of the authors’ propositions, and is missing some details. The proposition is original, but needs a better evaluation. The results are quite limited and simple, and, in my opinion, there is a need for more experiments and a better evaluation of the proposed method in order to demonstrate its real applicability and robustness. For example, the authors concentrate their efforts on demonstrating the reduction of false-positives, but they should discuss the overall evaluation of the performance of their method compared to other methods, when faced with different input data (images with different qualities and distortions). The comparison with the other techniques should be better detailed. In my opinion, this work should be improved with more experiments, results, and details.

Reviewer:  Fernando Osorio Review #: CR134420
Bookmark and Share
 
Pattern Analysis (I.5.2 ... )
 
 
Biology And Genetics (J.3 ... )
 
 
Feature Evaluation And Selection (I.5.2 ... )
 
 
Design Methodology (I.5.2 )
 
 
Segmentation (I.4.6 )
 
 
Life And Medical Sciences (J.3 )
 
Would you recommend this review?
yes
no
Other reviews under "Pattern Analysis": Date
Understanding data pattern processing
Inmon W., Osterfelt S., QED Information Sciences, Inc., Wellesley, MA, 1991. Type: Book (9780894353864)
Jun 1 1992
Parallel thinning with two-subiteration algorithms
Guo Z., Hall R. Communications of the ACM 32(3): 359-373, 1989. Type: Article
Jan 1 1990
A variable window approach to early vision
Boykov Y., Veksler O., Zabith R. IEEE Transactions on Pattern Analysis and Machine Intelligence 20(12): 1283-1294, 1998. Type: Article
Oct 1 1999
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy