Grabbing positions of sites are concordant, discordant and uninformative upon -scf/-scfl

[fkmendes]

Hello there!

As suggested by Rob, I'm posting to ask if it would be possible to implement a simple printing feature to -scf/-scfl calls. (Apologies if this post ended up in the wrong location of the forum; if so, maybe there's a way to move it?)

Basically what I wish I could obtain from IQ-TREE are the positions of concordant, discordant, and autapomorphic (fully parsimonially uninformative) in an alignment.

(At any rate, what a discordant site is discordant about is a different matter -- and I wonder how big the output file would be if all evaluated quartets could also be printed? But at a first moment, it'd be great to just know which sites fall into those 3 classes.)

I hope I was clear enough in my description above, please let me know if this makes sense.

Thanks a lot!
Fabio

[roblanf]

Hey fabio,

What exactly would you want the output to look like here. There are a LOT of options, and file sizes are likely to get huge. When planning it out, think about the general user too. What are the use cases, and what exact output formats would facilitate these the most?

Rob

[fkmendes]

Hi Rob,

Thanks for the quick reply!

At minimum, a file like this would be helpful:

concordant 1,3-8,11,13-23,...
discordant 2,9-10,12,...
uninf 14,...

In the above, for both concordant and discordant sites, we don't get which quartet that site is discordant for, only that it is. This would make the output be very very tiny.

A second, heftier output would be to print the quartet for which a site is concordant or discordant, and the terminal(s) that a site is autapomorphic for (if it is not invariant). I can see the concordant sites and uninformative sites not taking that space, but if the user wants information on discordant sites, the file can grow very large.

If we ignore discordant sites for a second, I'd imagine additional files like this:

concordant_quartets_list.txt
[file that lists quartets in the user-provided tree during -scf command, and some form of ID for each of them]

concordant.tsv:
[position] [quartet_ID]
1 quartet_ID
3 quartet_ID
4 quartet_ID
5 quartet_ID ... (where quartet_IDs come from concordant_quartets_list.txt)

uninf.tsv
[position] [autapomorphic_terminal or NA]
14 'sp1'
15 'sp1','sp15' (sp1 and sp15 would need to have different alleles here, otherwise this site would be discordant, I think?)
16 NA (use NA if site is invariant)

I don't know if what I'm describing above makes sense -- what do you think Rob?

Thanks a lot!