8 Mar 2012 mclblastline 12-068
mclblastline — a pipeline for clustering from BLAST files.
mclblastline [deblast options] [pipeline options] file-name
mcl has acquired the ability to proceed from label input as produced by mcxdeblast. This enables a very lightweight mechanism of generating clusterings from BLAST files. You might want to use this mechanism, documented in the mcl manual.
mclblastline used to require (given default parameters) the presence of the zoem macro processor to produce detailed output. This is no longer the case. By default mclblastline now creates a line-based tab-separated dump file. Zoem will be invoked when the --fmt-fancy option is supplied. In that case, zoem obviously has to be installed.
mclblastline wraps around mclpipeline. It supplies the --parser=app and --parser-tag=str options, setting them respectively to mcxdeblast and blast. This tells mclpipeline to use mcxdeblast as the parse script in its pipeline. The significance of the blast tag is that any mcxdeblast option can be passed through mclblastline and mclpipeline by inserting this tag into the option. For example, mcxdeblast accepts the --score=x option. When using mclblastline, you specify it as --blast-score=x. There are two exceptions to this rule, namely the --xi-dat=str and --xo-dat=str options. Refer to the mclpipeline manual for more information.
Additionally, all mclpipeline options are acceptable to mcxdeblast as well. The --whatif is useful for getting a feel for the pipeline. The --mcl-I=f inflation option and --mcl-scheme=i scheme index options are your basic means for respectively manipulating cluster granularity and allocating resources. Read the mcl manual entries for a description of the corresponding -I and -scheme mcl options.
The best advice is to glance over the mcxdeblast and mclpipeline options in order to get a feeling for which of those may come in handy for you. Then start experimenting. Use the --whatif option, it will tell you what would happen without actually doing it.
All mcxdeblast and mclpipeline options. mcxdeblast options must be passed using the mechanism described above.
This will use bit scores, sort cluster indices such that the corresponding labels are ordered alphabetically, ignore bit scores not exceeding 5, and use inflation value 2.5. In this case, the output clustering will be in the file named myblastfile.I25s2 (I25 identifying the inflation value and s2 identifying the resource scheme) and the formatted output will be in the file myblastfile.I25s2.fmt.
The first run prepares an input matrix to be read by mcl. In this case its file will be named myblastfile.sym. The subsequent runs use this matrix. CAVEAT there are some options that you need to repeat when executing such a resumed run. They are clearly marked in the mclpipeline manual - namely those options that affect names of (intermediate) files. Most importantly, this concerns the mclpipeline options that have prefix --xo or --xi. For example,
In this case, the matrix file will be named myblastfile.b.sym, and the --xo-dat options must be repeated in all runs so that the pipeline reconstructs the correct file name(s).
Stijn van Dongen
mcxdeblast, mclpipeline, mcxassemble.