The HORMONOMETER is a tool for evaluating any transcriptome response through the perspective of similar events that happen upon hormonal activation (Volodarsky et al, 2009). The relative rank for each hormone signature in the screened experiment is calculated based on comparing the transcriptome of the experiment to an indexed list of pre-calculated transcriptome responses established by published hormone treatments from Goda et al (2008). The gene identification list of all the hormone indices and their fold expression values can be downloaded from here. The comparison between the transcriptomes is done by a vector algebra algorithm that summarizes the up and down-regulated behavior of about 1,000 indexed genes for each hormone.
Input files require a ".csv" extension. The file should include the Probeset ID in column A, the locus (AGI) in column B, and then columns for p- value and fold change for each transcriptome of interest. The fold change is pre-calculated by the user by directly dividing the normalized results i.e. treated sample compared to a control transcriptome. If several experiments are to be scanned, the table should include a column of fold change and a column of p values for each of the experiments. P values are calculated by ANOVA. For example it is possible to calculate the fold changes and the p values utilizing the "PARTEK Genomics solution" (Downey, 2006). The processing of the data includes quantile normalization according to the RMA algorithm (Irizarry et al., 2003) and standard ANOVA model, treating each condition as a factor. An example file for 2 experiments is shown in Figure 1A. The fold change (Gfold) for 1 experiment is in row C and its p value is in row D. The second experiment is in rows E and F, respectively. A 'captions file' with ".csv" extension should be submitted as well: this file should have captions F P E in line 1 of columns A B and C respectively, and below the names of the columns of the fold change and p-value from the input file for each experiment. An example file for 2 experiments is shown In Figure 1B.
Figure 1: Example of data input to HORMONOMETER
You can download some test files below to test the system.
input data tables, captions (instructions).
After processing, the user receives a compressed (zip) file containing one or two files and a folder. The "analysis_result.csv" file lists the correlation ranks of the submitted experiments to each of the hormones. If several experiments are submitted, the output will include another file "analysis_result.pdf" showing a clustergram which arranges the experiments according to their similarity to each other in terms of the ranks received in the "HORMONOMETER". The folder contains a number of files listing all of the transcripts that were actually used in each particular query.
Figure 2: HORMONOMETER output of transcriptomes of wounded tissue harvested at different times after wounding. Data from Kilian et al. (2007) shows induction of rapid jasmonic signatures followed by latter ABA 1-6 h and latter ACC signatures 6-24 h. Note the repression in ABA signatures at 12-24 h