- Gene count files now has a triple index,
  the first is the rank of the original
  orthogroup, the second is the ID of the
  original orthogroup and the third is
  the ID of the matched OrthoFinder
  orthogroup. Rows are sorted by rank.

- The original GFs with the largest overlapping
  percentage on the OrthoFinder group is chosen
  between the two (or they can be also more than
  two).

- Now output shows not only the OrthoFinder GF
  name but also the name of the original top
  GF for which the match was found!

- I was so far using mock species, but now
  that I've generated the file I needed
  (results from OrthoFinder where the fagales
  clade was added altogether to update the
  37-species orthogroups) I can use the
  actual name of the fagales species in the clade

- New strategy: the small calde is added
  to the 37-orthofinder results altogether
- This new code parses the resulting OF
  orthogroups and 1) gets the equivalent
  of the original top 100 orthogroups; 2)
  gets those which at least one clade
  species has genes.
- Finally computes and outputs the gene
  counts for such families for the whole
  clade in a tab-seprataed DataFrame

- When original GF hits two OrthoFinder GFs,
  these two latter ones are merged and they
  are called with sorted label "GF1_GF2"

- Gene counts are printed to file with GF
  as index labels; GFs are sorted by
  descending gene count

- finally included cucumis melo and hordeum
  vulgare, so that I now have all the 37
  angiosperm species that were used in the past
  for the rec.ret. gene families
- ksrates code and main.nf run OrthoFinder
  from within the ksrates code
- ksrates also runs the analysis to find the top
  e.g. 2000 reciprocally retained gene families
  among the thosands of OrthoFinder orthogroups

- Allow accepting any OF orthogroup even if
  it contains many extra non-top genes.
- set option "test_run" to avoid generating
  files with incomplete data
- Generate filename variables

- For testing purposes, added switch parameter
  to switch between diamond and orthofinder.
- orthofinder itself is not yet run by ksrates
  because of dependency issues (diamond not
  found!). But the analysis of orthofinder
  results is performed to find the top 2000
  gene families in the OF orthogroups.
- Continues with classic Ks estimate and plots.

Triggering the reciprocally retained pipeline
relies on three variables in the main.nf
(top_gfs, rank_type and reciprocally_retained).
Before this fix,
if the expert configuration file was not
listing such three parameters, the NF pipeline
would have left them empty, raising issues
afterwards.
Now instead, if missing, they are initialized
with default values (2000, combined, no).

New ksrates command for multiple paralogs ks

See merge request !6

User-defined name/path for expert configuration

See merge request !4

Improve checkpoint for divergence colors provided in configuration file

See merge request !5

Perform lognormal mixture modeling on reciprocally retained paralogs

See merge request !3

Improve filtering steps in anchor clustering

See merge request !2

Integrate reciprocally retention workflow

See merge request !1