Chimera check shall optionally annotate possible sources

[westram]

• allow to specify "column-clusters" defined by quality-output of existing chimera check
  • e.g. specifying sth like "123,789" defines 3 column-types:
    • all columns with quality 1, 2 or 3
    • same for 7,8 and 9
    • rest (will be considered irrelevant)
  • for each column-type find column-clusters, specified by
    • max. percentage of "other qualities" inside the cluster and
    • min. length of cluster
• perform relative-search for each column-cluster (do position-specific search, maybe also global search if none found)
• for any found relative. calculate
  • mean degree of relatedness for each cluster type
  • mean sequence offset (only applies if global search is performed)
• sort found relatives "according to their probability to be a source of the suspected chimera"
  • annotate into a database field of the chimera