[Jluis]

Thank you for your swift answer, Simon.

My Java version is:

C:\Documents and Settings\jllavin>java -version

java version "1.6.0_31"
Java(TM) SE Runtime Environment (build 1.6.0_31-b05)
Java HotSpot(TM) Client VM (build 20.6-b01, mixed mode, sharing)


Using the command you provided, SeqMonk started (I've tested it up to 1200M), but doesn't seem to be able to open Bismark methylation extracted files, e.g. of those file format:

CHH_context_filename.fastq_bismark_pe.sam.txt

Maybe it is due to some format incompatibilities between Bismark/Seqmonk and not to the fact that Seqmonk is running on less memory than expected for that task (please excuse me if I skipped some post warning about this incompatibility issue before posting on this thread).

Anyway, thank you for your answer. It's always very nice to post in a forum where you get such fast & accurate answers.

[simonandrews]

Sorry this isn't working for you. It's not immediately obvious what the problem is from the error log you provided, it could be that there's a generic problem with java on your machine, or it could be a configuration problem.

Can you please try a couple of things for me. Firstly can you provide the output from running:

java -version

..just so I can see exactly which JRE we're working with. Can you also check that your machine really does have 3543MB RAM in it (I'd guess that's 4GB RAM, but with some taken up for graphics memory or the like).

Can you also try running:

java -cp "C:\Documents and Settings\jllavin\Desktop\SeqMonk" -Xmx500m uk.ac.babraham.SeqMonk.SeqMonkApplication

(not the change from 1500 to 500) to see if that produces anything. Hopefully we can get to the bottom of what's going wrong here.

[Jluis]

Hello everybody and specially to the seqmonk development team, to whom my question is aimed.

I can't start seqmonk in my windowsXP machine. I always get the same error (even tough I've tried to implement the solutions to fix the memory usage issue that where posted here).

Here it's the Running command used and Error mesage retrieved:

C:\Documents and Settings\jllavin\Desktop\SeqMonk>seqmonk -m 1000
(I also tried with seqmonk.exe in the command line and/or simply clicking on the file in windows).

Found 32-bit JVM, setting memory ceiling to 1500m
Physical memory installed is 3543
Amount of memory to use is 1500
Memcheck command is java -cp "C:\Documents and Settings\jllavin\Desktop\SeqMonk"
-Xmx1500m uk.ac.babraham.SeqMonk.SeqMonkApplication
Could not create the Java virtual machine.

Thanks in advance.

[fkrueger]

This is a really basic question, but which Bismark file is supposed to be imported into SeqMonk? According to the Seqmonk documentation it just says bismark file, which could be the SAM file from alignment or the txt files from the methylation extractor. I want to view methylation patterns. When I try importing the SAM file I get an error saying that the Bismark version numbers is missing from the file. If it is the txt files from the methylation extractor output I am not sure which one to import. I am interested in CpG methylation but there are 4 files related to that.

The current version of SeqMonk only supports the direct import of Bismark files in the 'vanilla' format, even though it is on the things to do list to update this for SAM format.

If your alignment results are in SAM format you need to go through the methylation extractor and import the result files via the generic text import (chr is col 3, start as well as end position are col 4, the "strand" (i.e. methylation state) is col 2).

You can either run the methylation extractor to get 1 combined file for all CpG calls (--comprehensive), or get 4 strand-specific files (OT, CTOT, OB and CTOB). If your experiment was directional you might want to delete CTOT and CTOB files and just use the OT (methylation on forward strand) and OB (methylation on reverse strand) files. If you are not interested in strand-specific methylation you can merge the two within SeqMonk by making a data group. Hope this helps.

[shawpa]

Which bismark file to import?

This is a really basic question, but which Bismark file is supposed to be imported into SeqMonk? According to the Seqmonk documentation it just says bismark file, which could be the SAM file from alignment or the txt files from the methylation extractor. I want to view methylation patterns. When I try importing the SAM file I get an error saying that the Bismark version numbers is missing from the file. If it is the txt files from the methylation extractor output I am not sure which one to import. I am interested in CpG methylation but there are 4 files related to that.

[simonandrews]

I've just released SeqMonk v0.21.0 onto our public server. I've been meaning to do a release for a while now as there have been a lot of additions to the program recently so there should be a lot of new stuff to play with. The main changes in this version are:

1. Full tracking information about parameters and quantitation in all filter results
2. A new Monte-Carlo statistical filter to allow the testing of a probe list to see how unusual it is in the context of the current probe set.
3. A new bisulphite sequencing quantitation pipeline.
4. A new hierarchical clustering plot which allows for rigorous clustering of probes
5. A new cis/trans quantitation method for HiC data.
6. Scatter plots and MA plots are now interactive - you can mouse over a point and see what it is, our double click it to move the display to that probe
7. A new option was added to allow the creation of multiple 4C datasets from a HiC dataset
8. Many improvements to the HiC heatmaps, including the addition of a clustering option
9. Improved statistics in the Intensity Difference filter making it much quicker and providing more results (due to a change from Bonferoni to Benjamini and Hochberg correction)
10. We fixed a bug in HiC import for datasets which contained warnings
11. We fixed a bug which would create invalid SVG files from some images
12. We fixed a bug which mis-scaled some sets of boxwhisker plots so they sometimes fell partially off screen

Please note that as part of this release the official project URL has changed to http://www.bioinformatics.babraham.a...jects/seqmonk/ (although the old address will continue to work). This change also means that we have had to update the seqmonk launchers in this release, so that if you have previously made your own launcher (or are using a very old .bat file launcher) then you will need to move to using the new launchers which come with seqmonk as the old ones will no longer work.

Please let me know if you find any problems in this new release.

[simonandrews]

All SeqMonk genomes are based on Ensembl releases. At the moment GRCm38 is so new it's still being processed on pre.ensembl.org and hasn't completed a full annotation cycle. As soon as the annotation is complete and this assembly makes it into the regular Ensembl system we will process this for SeqMonk and it will be available as a supported genome.

[Neuromancer]

Hey!

I have used SeqMonk for quite a while now. And I think it is really handy! As I'm working with mouse data I was wondering, whether to map my data again using the new GRCm38 (mm10) genome. But as far as I can see, SeqMonk does not support this, yet? Is there a possibility how I can use the new Assembly/Annotation in SeqMonk?

Thanks!

edit: Or is there maybe any reason why one should not use the new assembly yet?

[neurongs]

Hi Simons,

I realized that the plot gets the scale bar from the first plotted dataset. By changing the order in "set datatracks", I can fix it. Nevertheless, there is still a little difference in the median between the different representations of the same quantification data...

Thanks you in advance for your reply.

[neurongs]

Hi Simons,

First, I find seqmonk very interesting and I would like to thank you for the development of such an excellent tool.

I am analysing some ChIPseq datasets. I am a bit surprised since I found a slight difference between the medians reflected in the boxwhisker plot and those calculated on the report of the probeset (including unannotated probes). Do you have any possible explanation to this?

In addition, some times, the whiskers fall out of the represented scale and therefore the plot is incomplete.

Thank your in advance for your help and your time.

[simonandrews]

Hi Simon,
I tried to import the bed files as annotated track but Seqmonk could not recognize those files. I just import them as BED files, and quantify by using "RPKM calculation for RNA-Seq data". Then I filter the data by intensity difference with p=0.05 cutoff (normal vs. patients) and save the feature report. To search for splice variant, I have to open the bed files in IGV, go through the genes in the feature report one by one. Do you think there might be better solution than this? Thank you!

Are these BED files "multi-location" BED files by any chance? If so then SeqMonk's BED parser won't recognise them. We did look at putting in support for them, but even the people who made the format were saying that they were not recommended for use and people should switch to GFFv3 or GTF.

Is there any way you could let me have a copy of your results for one experiment so I can actually see what you're working with. It's difficult to offer more useful suggestions when I can't actually see the data.

[mediator]

Hi Simon,
I tried to import the bed files as annotated track but Seqmonk could not recognize those files. I just import them as BED files, and quantify by using "RPKM calculation for RNA-Seq data". Then I filter the data by intensity difference with p=0.05 cutoff (normal vs. patients) and save the feature report. To search for splice variant, I have to open the bed files in IGV, go through the genes in the feature report one by one. Do you think there might be better solution than this? Thank you!

[simonandrews]

Hi Simon,
Thanks for the help! Will try that.

Let me know if it works out. I can take a better look if this approach turns out not to be feasible.

[mediator]

Hi Simon,
Thanks for the help! Will try that.

[simonandrews]

Hi Simon,
For bed file (generated by Scripture, from RNA-Seq data), which quantification pipeline would you recommend? I am trying to compare bed files between patients and healthy controls in order to find splice variants unique to patients. Thank you!

I've not used scripture before, but looking at the documentation it looks like the data you get out of scripture is probably more processed than you'd want to put into SeqMonk as a data track. We'd normally import the output of Tophat into the program, either importing the spliced exonic reads, or the introns depending on what we were looking for.

From what I can see scripture tries to create assembled transcripts from your raw data, so I guess the best way to handle this would be to import it as an annotation track rather than a data track. If the features it produces are spliced then you'd need to import them as GTF or GFFv3 files since none of the other annotation formats supported by SeqMonk can handle multi-location features.

Once you have these elements in place then you could quantitate the various scripture transcripts in your datasets and then compare these. You could use the standard RNA-Seq quantitation pipeline and follow the basic RNA-Seq methodology (I'm actually in the process of producing an improved RNA-Seq guide since we have a pretty solid way of dealing with this data now).