simseq-0.0: idea.txt

Simple tool to generate simulated ESTs.

Must support (minimum, generalize?):

     sanger: initial medium/high error rate,
	     stable low error rate (mostly subst?)
	     slowly increasing to one from pos 750-900

     454: medium error rate, mostly ins/del for monomers

     Solexa: any idea of error rate?

Generalized:
     general error models, training on data sets?
     HMM-like?

** PLAN **

  primer: determines starting point (initial state) from sequences
  transcribe: unfold new sequence, while modifying state
  error model: depends on state
  terminator: ends transcript    

Specialized: homopolymers for 454 (must be part of state)


  Terminator = Distribution
  Error Model = Distribution + Mutator
  Primer = ...?

Mutators can be 'ins(c,p)', 'del(p)', 'subst(c,p) etc.
Or should the distribution multiplier apply already to distrib?

-k