%0 Journal Article
%T Correction: Three-parameter lognormal distribution ubiquitously found in cDNA microarray data and its application to parametric data treatment
%A Tomokazu Konishi
%J BMC Bioinformatics
%D 2004
%I BioMed Central
%R 10.1186/1471-2105-5-82
%X f(ri) = [k/{(2¦Ð)1/2 ¦Ò(ri - ¦Ã)}] exp [-{log(ri - ¦Ã) - ¦Ì}2/2¦Ò2] for ri > ¦Ã,where k is a compensation constant (k = loge = 0.4343), ¦Ò and ¦Ì are the shape and scale parameters for log(ri - ¦Ã), respectively.The threshold parameter, ¦Ã, was found through trial and improvement calculation processes; in the trial, the distribution of log(ri - ¦Ã) was checked by normal probability plotting, and the value that gave the best fit to the model was selected for ¦Ã. The fitness was evaluated by the sum of absolute differences between the model and log(ri - ¦Ã), within the interquartile range of data. The parameter ¦Ì was found as the median of log(ri - ¦Ã), and the parameter ¦Ò was found from the interquartile range of log(ri - ¦Ã); these are known as robust alternatives for the arithmetic mean and standard deviation, respectively. Parameters ¦Ì and ¦Ò were found for each data grid, a group of data for DNA spots that were printed by an identical pin in order to avoid divergences caused by pin-based differences. Z-normalization was carried out for each datum asZri = {log(ri - ¦Ã) - ¦Ì}/¦Ò.Intensity data (ri) less than ¦Ã were treated as "data not detected", since such data might contain negative noise larger than the signal (see Results).
%U http://www.biomedcentral.com/1471-2105/5/82