Comparisons of BGC

[HenrikSpiegel]

Documentation of how we compared BGCs

Attempted solutions:

• Needleman wunsch

  • Did not take into account length difference
  • could not take into account that the seqeunce of genes were not preserved.

• ANI

  • Not suitable for large number of pairwise comparisons.

• fastANI

  • Very fast, but very inaccurate for our application
  • Was entirely dependent on the input parameters and results would very very heavily.

• Blastn approach

  • Requires post-processing
  • -subject_besthit flag appears to perform some overlap collapsing within subject-query. (Need documentation).
  • Similarity metric is also needed to obtain a single score for each pw-comparison.

Figures required:

Create a blast illustration looking like the below: should be doable for the 16v16 illustration.