RELEASE NOTES FOR ChIPMonk V1.2.3 --------------------------------- ChIPMonk v1.2.3 is a bugfix release which corrects a couple of problems which prevented the program from loading some of the newer genome assemblies which had become available. All of the genomes in the repository should now be able to be read by ChIPMonk. In addition, the OSX version of ChIPMonk is now able to run under a 64-bit java virtual machine, so users with the appropriate hardware can now increase their memory usage above the 1.5GB effective limit on 32 bit machines (this was always possible on 64 bit windows machines). RELEASE NOTES FOR ChIPMonk V1.2.2 --------------------------------- This is a minor bugfix release. It corrects a problem with the annotation of features in the annotated list report. In older versions of the program this report labelled upstream features as downstream and vice versa. The options set when generating the report were used correctly and the results were correct but just annotated with the wrong label. RELEASE NOTES FOR ChIPMonk V1.2.1 --------------------------------- Release v1.2.1 is a bugfix release which fixes a couple of common issues which have been reported. Most of the important fixes are in the genome parsing code. Many people were finding that ChIPMonk would freeze whilst trying to download or parse genome files. This release includes much better error handling code which will either cope with these problems or at least produce an informative error message. There have also been numerous smaller bugfixes within the interface. There are no major functional additions in this release and there are no known regressions from previous releases. We would recommend that all users update to this latest version of the software. RELEASE NOTES FOR ChIPMonk V1.2 ------------------------------- Text Import ----------- ChipMonk now supports the importing of data from any plain text format. This is a new feature and although it has been tested there may still be issues with some formats we've not seen before. The import occurs in two stages 1) The importing of probe data 2) The importing of array data for the probes In the first step your file needs to supply the probe id, chromosome and position (either a central position or a start and end position). If this information is present in all of your data files then just pick one of them to use to imort it. In the second step your file will need to contain probe ids which match those imported in the first step and a series of experiment and control values for each probe. You can have more than one set of experiment/control values in a single file. Please report any problems you have with this new feature, preferably including a copy of any file which proved problematic. Import Warnings --------------- When ChIPMonk loads data there are a series of non-fatal errors which it can encounter. In previous versions of the program these were not shown so you didn't know if there were problems with a subset of your data. The program will now list the warnings after finishing any data importing. Often these warnings are harmless, but you should read them to make sure there isn't a problem with your data. The most common warnings you're likely to get are: 1) The position of a probe is off the end of its chromosome. If you get a lot of these warnings it is usually an indication that you have used the wrong genome assembly for your experiment. This would mean that all of your probes will have been incorrectly positioned (although maybe not by much). If you just get a small number of these errors it's probably not something to worry about. 2) A chromosome could not be found. This warning occurs when the name of the chromosome in your data doesn't exactly match the name in the assembly. The most common time for this to occur is for the mitochondrion which can be named M, Mt, Mito or all sorts of things. A worse situation is if you have a mix of arabic and roman numerals for your chromosome names (ie 2 vs II) in which case none of the chromosomes will be found. We hope to have a work round for this problem in the next release of ChIPMonk, but in the mean time please contact us if you're having this problem and we can tell you how to fix it manually. RELEASE NOTES FOR ChIPMonk V1.1 ------------------------------- Statistical bug fixes --------------------- v1.1 fixes a few bugs in the calculation of statistical parameters which were present in v1.0. These changes will mean that p-values calculated using older versions will be different from those in the newer version. P-values should not change wildly, and the ranking of probes will not be different but we would advise that you recalculate any important filters using the updated version of the program. The bugs were in the Student t-test distribution calculation where a 1-tailed rather than a 2-tailed p-value was being calculated. This had the effect of halving every p-value. Also the Bejamini & Hochberg correction was being applied incorrectly such that genes would appear more significant than they should. Updated Genome Annotations -------------------------- Alongside the updates the the ChIPMonk program we have also released updates to the Genome annotation files which ChIPMonk uses. These have been improved to add more useful gene names and descriptions and to include GeneOntology mappings and descriptions wherever possible. In order to take advantage of these updated annotations you will need to re-import the updated files into your ChIPMonk installation. The easiest way to do this is to select File > New Project > Import New and then select the genomes you are using (you don't actually have to go on and create a new project). This will overwrite your existing genome files and you should see the updated annotation. As part of the enhancement of the genome files we have had to rename some of the existing mouse assemblies to bring their names into line with those used by Ensembl. The only affected genomes are Mouse assemblies called NCBI## (where ## is a number). The updated versions of these genomes are called NCBIM## (note the extra M). If you want to use these new genomes you must edit your .cmk files in a text editor. The second line of each file should say: Genome Mus musculus NCBI36 You should change this to read: Genome Mus musculus NCBIM36 When you next load the data ChIPMonk should load the new version of the annotation. Changes to semantics of report annotation ----------------------------------------- When I re-read the text on the annotated probe and probe group reports I realised that the program wasn't actually doing what the description in the dialog boxes described. Specifically the terms upstream and downstream were being used the wrong way round. In the code upstream and downsream were used to refer to the position of the probe, but in the text in the dialog they apply to the position of the feature. I've changed the code so that it now does what the text says it should, but be aware that this is a change in behaviour from the previous versions. Just to be clear. If you wanted to annotate probes which appeared in the promoter of a gene you would annotate them with DOWNSTREAM gene features (since the feature will be downstream of the probe). Some Nimblegen files import incorrectly --------------------------------------- We have been sent some Nimblegen files which ChIPMonk will import incorrectly, mapping some probes to the wrong chromosome. The reason for this is that one of the versions of Nimblegens software names its probes in such a way that it is impossible to tell which chromosome they come from. To give you the gory details; ChIPMonk uses the all_pair.txt file (which is now often named after the experiment rather than all_pair.txt, but is otherwise the same file) because it is the only single file which provides all of the information needed to import your data. Within this file we have to extract the chromsome and position for each probe from the probe name since this is the only field which contains all of this information for all of the arrays we've seen. Older arrays contained probes with names like CHR1P004094210. From this you got Chr 1 and position 4094210. Easy. In later arrays we started to see names like CHR0100P004094210 where the first two characters after CHR indicated the chromosome number, padded to the left with zeros. More difficult for very little benefit, but still OK. Unfortunately at least one array we've seen had names like CHR1000P004094210, where the probe was on Chromosome 1. In these arrays we're stuffed as there's absolutely no way to tell from the probe name alone which probes are on chr 1 and which are on chr 10 (or 2 vs 20 etc). If you have an array which is affected in this way please tell Nimblegen and ask them to pad their chromosome names to the LEFT! In some cases we can fix these files if they have the chromosome name in the sequence name field of the all pair file. If you're stuck with this kind of data then please contact us and we'll see if we can sort it out for you.