TODO

IMPORTANT
---------

- consensus output (sto,aln) and compatibilety with bioperl/biopython

- Split Homfam refs in two. Use one part as background HMM, the other parts
  for benchmarking

- Implement meta flags:
  accurate: --iterations 3
  default:  --mbed --iterations 1
  fast: --mbed
  and for more than 10k sequences: --mbed --mbed-iter

- SSE instructions for hhalign (little use in ClustalO frontend; DD)
  /
  Patch new code which already contains SSE instructions

  also fix automake/configure then

- Multi-HMMs; also for Pfam+iteration (FS)

- Show degradation of alignment quality when using x reference sequences
  added to y random/false sequences
  (Lio Pachter)

- Seed pre-alignment with M-Coffee, MSAProbs, ...

- GUI/API:
  Will have to catch/override exits from source.
  find . -name \*.c -or -name \*.cpp -or -name \*.h | xargs grep 'exit('
  Also best to allow for user override of
  void Fatal(char *msg, ...);
  void Error(char *msg, ...);
  void Warn(char *msg, ...);
  void Info(int level, char *msg, ...);


- Soeding: DNA/RNA alignment incl. reading of nucleotide HMMs

- Automatic HMM-selection/search/download for input

- Structure input: Psipred predictions are apparently part of their
  hhms and should therefore be ready to use (part automatic
  HMM-selection)