search

VIB-PSB / ksrates

ksrates is a tool to position whole-genome duplications relative to speciation events using substitution-rate-adjusted mixed paralog-ortholog Ks distributions.
https://ksrates.readthedocs.io
GNU General Public License v3.0
15 stars 9 forks source link

Difficulty in Output Plots Interpretation #59

Closed abcbot123 closed 5 months ago

abcbot123 commented 5 months ago

Hi Cecilia, I have two question regarding my outputs,

1) I didn't get this output "mixed_oxalis_lmm_colinearity.pdf" mentioned in your tutorial. (Ran with expert configuration file). But I got this output file "mixed_species_lmm_paranome.pdf" Can I know the reason for that?

2) This is regarding the main output file "mixed_oxalis_anchor_clusters.pdf". In this plot, I got just one peak. In the literature, I know that ancient γ event occur. How can I know that one peak is for the ancient γ event or it is species specific WGD? Here I attached the "mixed_oxalis_anchor_clusters.pdf" and "mixed_oxalis_adjusted.pdf" files. Could you please help me to interpret these results. mixed_oxalis_anchor_clusters.pdf mixed_oxalis_adjusted.pdf

Cecilia-Sensalari commented 5 months ago

Hi!

Thanks for reaching out. Let's see:

  1. To get the LMM for anchors you have to set colinearity = yes (paranome doesn't matter) and extra_paralogs_analyses_methods = yes. Could you double check that? _UPDATE: at present time, documentation mentions mixed_species_lmm_colinearity.pdf as output filename, while the actual filename is mixed_species_lmm_anchors.pdf. To be corrected in the docs._

  2. The clustering in mixed_oxalis_anchor_clusters.pdf has been already "cleaned up" of clusters that are likely to be spurious, namely not an actual WGM signal. This can happen when they are very tiny or else widespread over high Ks. In your case, I bet that the ancient gamma peak appears quite flat and widespread around 2.5Ks; its cluster must have been unrightfully removed by the cleanup. I suggest you use another PDF file without filtering, where you should still see the gamma cluster: mixed_oxalis_anchor_clusters_unfiltered.pdf, located within the paralogs_analyses directory; note that you might also see some other spurious clusters that were instead rightfully removed.

  3. I'll add the last answer as a separate bullet point. Even without having the gamma cluster visible in the plot, the blue cluster "a" is not gamma because it is not shared between Oxalis and Arabidopsis and between Oxalis and Populus: the "a" WGM occurred more recently (0.76 Ks) then the divergence with any of them (1.73 Ks). However, you should also be careful in claiming that it is specific to your focal species, as perhaps there are other closely-related species that do share WGM "a" with it, they are just not included in your plot. A check in literature might help?

Hope this helps, Cecilia

abcbot123 commented 5 months ago

Thanks so much Cecilia. This is really helpful.

Cecilia-Sensalari commented 5 months ago

Alright! I'll close this issue then. Cheers, Cecilia