AB SOLiD Bioscope

[rdeborja]

Is anyone using/testing Bioscope as a replacement for corona lite and the whole transcriptome pipeline? I've recently installed it on our cluster and was curious to find other opinions/experiences with it.

[drio]

I have been using it for re-sequencing (still testing). BS bundles a bunch of different experiment, WT among them. I would say, download the software and start by
running the examples that come with it. Once you have it up and running modify it
to work with your data and test it out.
ABi supports SGE and PBS. THe installation in non-root mode is not extremely invasive so I would suggest you start by that.

Let us know how it goes.

[westerman]

Bioscope seems ... fragile ... at the moment. Fragile in the sense that I can run all of the test examples just fine and they seemly give good results. But when I modify the scripts to work with my data then things fall apart.

[drio]

Can you be more specific? What do you mean fall apart?

[westerman]

Bioscope has, in theory, a very nice workflow where you specify a 'plan' of modules to use for a given project. As an example the 'plan' may call the 'quality value filter' module, then the 'mapping' module, then in parallel the 'mapping statistics' module plus the 'GFF' module. And so on. Each module has their own 'ini' (initialization) file which in addition to per-module commands can read in a global ini file and a per-project ini file. Once you have a 'plan' set up then you just hand it off to the workflow schedule and it runs the project.

All very nice. Except ... the ini files do not pass information between each other. Nor do they use consistent parameter names. Some of the parameters are used but undocumented. The modules make unwarranted assumptions on what the previous module has done. A common example involves file names where the output file name of a module may not be in the format that the next module can understand as its input file name. Ditto with other parameters. I set a parameter in the per-project ini file and expect it to filter down through the modules -- which it does to the most part but I will often find at least one module which does not accept the parameter.

That is what I mean by 'fragile' and 'fall apart'. Touch or change one small part of the plan and the pipeline fails.

Of course this is version 1.0 of Bioscope. The software is really only used by a handful of people (as compared to, say, Microsoft Word). Thus we should expect that we will be de-facto beta testers and will encounter rough spots. Lifetech/ABI support is very responsive in trying to fix problems that I find.

[drio]

Bioscope has, in theory, a very nice workflow where you specify a 'plan' of modules to use for a given project. As an example the 'plan' may call the 'quality value filter' module, then the 'mapping' module, then in parallel the 'mapping statistics' module plus the 'GFF' module. And so on. Each module has their own 'ini' (initialization) file which in addition to per-module commands can read in a global ini file and a per-project ini file. Once you have a 'plan' set up then you just hand it off to the workflow schedule and it runs the project.

All very nice. Except ... the ini files do not pass information between each other. Nor do they use consistent parameter names. Some of the parameters are used but undocumented. The modules make unwarranted assumptions on what the previous module has done. A common example involves file names where the output file name of a module may not be in the format that the next module can understand as its input file name. Ditto with other parameters. I set a parameter in the per-project ini file and expect it to filter down through the modules -- which it does to the most part but I will often find at least one module which does not accept the parameter.

That is what I mean by 'fragile' and 'fall apart'. Touch or change one small part of the plan and the pipeline fails.

Of course this is version 1.0 of Bioscope. The software is really only used by a handful of people (as compared to, say, Microsoft Word). Thus we should expect that we will be de-facto beta testers and will encounter rough spots. Lifetech/ABI support is very responsive in trying to fix problems that I find.

Thanks for sharing,

My five cents:

PROs:

+ dramatic improvement in terms of running time compared with CL
+ increase of sensitivity with same specificity.
+ Much more resource efficient both IO and CPU (it is multithreaded now)
+ Easier to start analysis (at least compared to corona lite)

CONs:

+ Still to many unnecessary files being generated
+ BAM is not the standard format to drop the alignments. Valuable
CPU cycles and I/O bandwidth wasted in postprocessing.
+ Changes in the reporting stats don't match the old corona lite. They mainly
report uniquely mapped reads.

[skblazer]

I have been using it for re-sequencing (still testing). BS bundles a bunch of different experiment, WT among them. I would say, download the software and start by
running the examples that come with it. Once you have it up and running modify it
to work with your data and test it out.
ABi supports SGE and PBS. THe installation in non-root mode is not extremely invasive so I would suggest you start by that.

Let us know how it goes.

Excuse me, where can I download Bioscope

[drio]

Excuse me, where can I download Bioscope

http://solidsoftwaretools.com

[skblazer]

http://solidsoftwaretools.com

I am sorry, but I cannot find Bioscope here.

[westerman]

I think that ABI/LifeTech are still just releasing the bioscope software on an 'as-need' basis until such time as they have it in releasable form. The last public link that have at the web site is for SAET. I do not see a public mention of bioscope.

[jaldrich]

Sorry a little late to this discussion, but has anyone been able to get Bioscope to run on their own data? (that is the CL version BioScope-1.0.1-42)

[westerman]

Yes. I can run fragment to diBayes calls, pairing to diBayes calls and most recently the whole transcriptome calling. All command line. It has been a bear to get running smoothly since the assumptions that the various pipelines run under seem to be different.

[jaldrich]

That is great news that you were able to get it running on your own data. Thanks for the quick post. I am in the process of getting up running on our data too. Just wanted to know if someone has been successful.

[clariet]

Just saw this post. We were able to use a Whole-transcriptome pipeline of BioScope (1.0.1-42) on a RNA-seq dataset. And a note about its mapping statistics. I confirmed with their specialists that the current version of BS has bug on those numbers. so it will be fixed in next release, hopefully very soon.

We have a feeling that a large proportion of reads are wasted for SOLiD data compared to Solexa. For example, for a current chip-seq dataset, we have seen a average of 80M reads generated for a sample (quad). However, after filtering of low quality alignment and non-unique hits, only ~4% of reads could be used for further peak detection. Has anyone have similar experience? Does this sound normal?

[cchu]

Yes. I can run fragment to diBayes calls, pairing to diBayes calls and most recently the whole transcriptome calling. All command line. It has been a bear to get running smoothly since the assumptions that the various pipelines run under seem to be different.

How many nodes, processors and memory you have for the offline cluster? We installed bioscope on the cluster and got it run using the example data but failed when we moved to test our own data with whole transcriptome pipeline. Dose it need a lot of memory to run?