They were later transferred to secure tank facilities where they lived for more than four months

and study into a single spreadsheet. Further, it will insert the correct If the “Make statistics” button is activated, ADE will perform a statistical analysis of the data and provide a combined p-value for a gene being up- or down-regulated in all studies in the merged output file. As ADE will not overwrite existing files, it is possible to perform statistical analysis after an initial run, allowing the user to delete fractions or sort the merged output file before performing statistics must remain unchanged). ADE will first calculate p-values for each gene in the respective studies by performing a one-tailed Student’s t-test. To merge p-values, we included Stouffer’s Ztransform method and Fischer’s method. For Stouffer’s Z-transform method, p-values are first transformed to Z scores, with Zi = – 1, being the standard normal cumulative distribution function. Z-Scores receive signs according to the gene being up- or down-regulated and summed to an overall Z-Score, pffiffiffiffi with Z s N Z i = N, N being the number of Z-scores. i1 Z-scores are not weighted in this approach. Zs is finally transformed to a combined two-sided p-value with ps = 2. Fisher’s method uses 2 -2k In pi, i1 2k where 2 is a chi-squared distribution with 2k degrees 2k Kurtenbach et al. BMC Research Notes 2013, 6:496 http://www.biomedcentral.com/1756-0500/6/496 Page 4 of 7 of freedom, k being the number of p-values. We found Stouffer’s method to be more restrictive, but other statistical approaches may also be considered. Findings Hundreds of G protein-coupled receptors are known to be expressed in the heart, some of which linked to heart disease formation. To demonstrate ADE function, we downloaded a list containing 288 class A GPCRs from the International Union of Basic and Clinical Pharmacology database website and expression data from 12 studies comparing normal and diseased human hearts from the GEO database. Datasets contained data for various disease groups. Downloading and organizing all files for ADE will take ~15 min per study initially, but once prepared files can be stored and reused for ADE runs. After preparation, ADE took ~5 minutes to extract and refine the data for all GPCRs on a standard desktop computer. purchase CJ-023423 groups with less than 5 samples and genes covered by less then five studies were deleted manually from the merged output. Further, genes where <75% of the studies agreed on PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19799341 up-or down-regulation were excluded. P-values were calculated for the remaining 25 groups and 14 different heart disease conditions. It may be considered to separate/exclude certain disease groups, but here we exemplarily performed statistics on the complete data. Statistical analysis took ~10 minutes to process. The analysis reported 43 class A GPCRs as differentially expressed in diseased human hearts with a p-value < 0.001, according to Stouffer's Z-transform method. Data was combined from 12 studies, containing 25 groups and 14 different cardiomyopathies: human arrhythmogenic right ventricular cardiomyopathy, dilated cardiomyopathy, doxorubicin induced cardiomyopathy, fetal cardiomyopathy, valvular cardiomyopathy, hypertrophic cardiomyopathy, and ischemic cardiomyopathy. p-value represents combined p-values using Stouffer's Z-transform method. +/- indicates up- or down-regulation reported by > 75% of the studies. For example, we found a highly significant downregulation of the beta 1 adrenergic receptor in 19/21 groups, which is well known for its crucial role in heart fun