add pos & neg DFE from Rodrigues et al (PosNeg_R24 in HomSap)

[petrelharp]

Motivation: we'd like to have a DFE containing positive mutations that aren't all at one fixed value of s.

Rodrigues, Kern & Ralph doesn't exactly infer such a DFE, in the sense of using a validated inference procedure, but does explore parameter space and find a best-fitting DFE - so, I think it counts, and since it is simulation-based, we know it produces roughly reasonable simulations. The targets of inference are only the proportions of positive and negative selection; the negative part of the DFE is from Castellano et al, which I also cite here; the positive part is based on "expert knowledge" (i.e., something that seems reasonable).

So, in the popsim spirit of "best guess based on current state of knowledge", I think this is a good one to put in. Other opinions welcome (esp if they propose a better alternative such DFE?).

Note that the method uses all the great apes, and so could reasonably be applied to any of them.

[petrelharp]

(Actually I was hoping @mufernando would QC this - I'll get a code review from someone else?)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.85%. Comparing base (6218f3d) to head (025da27).
Report is 11 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff            @@
##             main    #1657    +/-   ##
========================================
  Coverage   99.84%   99.85%            
========================================
  Files         132      136     +4     
  Lines        4584     4690   +106     
  Branches      467      470     +3     
========================================
+ Hits         4577     4683   +106     
  Misses          3        3            
  Partials        4        4            

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[petrelharp]

I've removed you from the assignee list, @mufernando, since I'm hoping you'll QC this? Say if you are actually going to review this PR instead.

[mufernando]

I will QC this instead Get Outlook for iOS<https://aka.ms/o0ukef>

…

________________________________ From: Peter Ralph ***@***.***> Sent: Tuesday, January 21, 2025 12:57:30 PM To: popsim-consortium/stdpopsim ***@***.***> Cc: Murillo F. Rodrigues ***@***.***>; Mention ***@***.***> Subject: Re: [popsim-consortium/stdpopsim] add pos & neg DFE from Rodrigues et al (PosNeg_R24 in HomSap) (PR #1657) I've removed you from the assignee list, @mufernando<https://github.com/mufernando>, since I'm hoping you'll QC this? Say if you are actually going to review this PR instead. — Reply to this email directly, view it on GitHub<#1657 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ACL5HFA5YOEHECD6CDMJB6T2L2YDVAVCNFSM6AAAAABVO3YKFOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMMBVG4ZDEOJRGM>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[silastittes]

Did the following locally with the PR

import stdpopsim
import numpy as np

species = stdpopsim.get_species("HomSap")
dfe = species.get_dfe("PosNeg_R24")
contig = species.get_contig(length=30000)
model = species.get_demographic_model("OutOfAfrica_3G09")
samples = {"YRI": 50, "CEU": 50, "CHB": 50}

gene_interval = np.array([[10000, 20000]])
contig.add_dfe(intervals=gene_interval, DFE=dfe)

engine = stdpopsim.get_engine("slim")
ts = engine.simulate(
    model,
    contig,
    samples,
    seed=236,
    slim_scaling_factor=10,
    slim_burn_in=10,
)
print(ts)
print(ts.diversity())
print(ts.allele_frequency_spectrum())

╔═══════════════════════════╗
║TreeSequence               ║
╠═══════════════╤═══════════╣
║Trees          │         61║
╟───────────────┼───────────╢
║Sequence Length│      30000║
╟───────────────┼───────────╢
║Time Units     │generations║
╟───────────────┼───────────╢
║Sample Nodes   │        300║
╟───────────────┼───────────╢
║Total Size     │  164.2 KiB║
╚═══════════════╧═══════════╝
╔═══════════╤════╤════════╤════════════╗
║Table      │Rows│Size    │Has Metadata║
╠═══════════╪════╪════════╪════════════╣
║Edges      │ 819│25.6 KiB│          No║
╟───────────┼────┼────────┼────────────╢
║Individuals│ 150│16.4 KiB│         Yes║
╟───────────┼────┼────────┼────────────╢
║Migrations │   0│ 8 Bytes│          No║
╟───────────┼────┼────────┼────────────╢
║Mutations  │ 258│14.9 KiB│         Yes║
╟───────────┼────┼────────┼────────────╢
║Nodes      │ 639│24.4 KiB│         Yes║
...
╟───────────┼────┼────────┼────────────╢
║Sites      │ 255│ 6.2 KiB│          No║
╚═══════════╧════╧════════╧════════════╝


0.00010336900780379047
[0.00000000e+00 1.30000000e-03 1.66666667e-04 1.33333333e-04
 0.00000000e+00 6.66666667e-05 6.66666667e-05 0.00000000e+00
 6.66666667e-05 3.33333333e-05 6.66666667e-05 1.33333333e-04 
...
 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00
 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00
 0.00000000e+00 0.00000000e+00 0.00000000e+00 0.00000000e+00
 0.00000000e+00]

The rationale makes sense and all looks good to me!

[silastittes]

All seems good!