Primary source of Inchworm Documentation is the website:

    http://inchworm.sf.net

Basic installation instructions and usage are as follows:

Download the latest release from sourceforge.  

If you want the bleeding edge, pull it from SVN at your own risk:

   svn co https://inchworm.svn.sourceforge.net/svnroot/inchworm inchworm



** Installing Inchworm **

Hopefully, all you should need to do is this:

    ./configure --prefix=`pwd`

    make

    make install

and 'inchworm' will appear as ./bin/inchworm


Note: if you prefer to minimize memory allocations, you can enable the use of Google's sparse-hash like so:

   ./configure --prefix=`pwd`  CPPFLAGS=-D__GOOGLE__

The inchworm will run slower (~3x) but consume less memory.


** Running Inchworm **

Inchworm will assemble RNA-Seq data from a fasta-formatted data set.  Inchworm builds a debruijn graph and traversed the graph to report sequence assemblies.

By default, inchworm expects that the data are strand-specific and sense-oriented.

Run like so:

      inchworm --reads $fasta_file  --run_inchworm  > assemblies.SS.fasta


If you do not have strand-specific data, run in double-stranded mode:

     inchworm --reads $fasta_file --run_inchworm --DS > assemblies.DS.fasta


A useful option to include is:  --monitor 1 
which will report the status of parsing the reads while building the kmer graph.


There is no 'paired read' mode currently in inchworm.  It treats each read independently. If you have pairs, just include them all in a single fasta file with the sequences properly oriented for strand-specificity (sense strand).


The assembly sequences output by inchworm are formatted like so:

>a1;123

a1 corresponds to 'assembly 1', and 123 corresponds to an average kmer coverage (ie. read coverage) for the assembly.


By default, a Kmer value of 25 is used.  You can change this with the option -K


Questions, comments, etc?  Contact Brian Haas  bhaas@broadinstitute.org