FlowSim is a simulator for pyrosequencing.

It currently supports Roche 454 sequencing, and attempts to model
accurately the flowgram generation, including base and quality
calling.  The output is written as an SFF file.

Currently, only the original GS20 generation is supported, but the
newer FLX and Titanium generations should be easy to add, as soon as
we have reasonably accurate statistical models for the flow
distributions, and a better idea of how to do quality calling.

  Installation

The quick and dirty way to acquire flowsim is to download the statically
linked Linux executable from http://malde.org/~ketil/biohaskelll/linux_binaries/
If your system is similar enough to mine, this should work - but
please read the next paragraph as well!

Only slighltly more involved, and a lot less dirty, is to install the
cabal-install package which provides the 'cabal' utility.  You will
also need a working GHC (which probably includes MINGW or something on
Windows). This should work for most current operating systems, and you
only need to run the command 'cabal install flowsim' to download the
source for FlowSim and its dependencies from HackageDB, compile,
and install them.  Or any other program written in Haskell.

If you want the cutting edge, you can get the latest sources using darcs:
'darcs get http://malde.org/~ketil/biohaskell/flowsim'.  If you have everything
else installed (i.e. the GHC compiler, and various libraries), you should be
able to compile and install flowsim by typing (in the flowsim directory):

     chmod +x Setup.hs
     ./Setup.hs configure
     ./Setup.hs build
     ./Setup.hs install

You probably want to check out the various options to these first.  If
configuration fails with missing dependencies, you need to install these,
either using 'cabal install' or manually downloading them from
HackageDB (http://hackage.haskell.org/) and installing them.

  Using FlowSim

Usage is, in brief:

   clonesim [-c #] [<input.fasta>] | flowsim [-G gen] [-o <output.sff>]

Clonesim¹ is used to generate a set of clones from an input file
(basically generating fragments of random lengths and orientations,
from random locations in the input), and flowsim simulates
pyrosequencing process of the clones.  The 'gen' parameter may
currently be GS20, Titanium or EmpTitanium.  Both clonesim and flowsim
may be given the --help option for more extensive usage information.

So, to generate an SFF read consisting of 100K simulated shotgun reads
from genome.fasta using the empirical flow distributions could be done
thusly:

    clonesim -c 100000 genome.fasta | flowsim -G Emp -o sim.sff

If you are interested in using this, or in further development, please
contact me at ketil dot malde at imr dot no.

Note that clonesim was added (or rather, factored out from flowsim) in
version 0.2.6,