Feature idea: new methods to species regarding substitutions

[vsudbrack]

I was wondering if it could be useful to add methods for species that handle substitutions similarly to how genome (or individual) methods handle mutations. For instance:

sim.substitutionsOfType(mutType id);
sim.sumOfSubstitutionsOfType(mutType id);
sim.countOfSubstitutionsOfType(mutType id);

I think any simulation of quantitative traits could benefit, as the trait of an individual of species sim based on mutations of type m1 (e.g. "n" with mean 0 and given variance) could be readily written as

phenos = sim.sumOfSubstitutionsOfType(m1) + inds.sumOfMutationsOfType(m1);

A few recipes that could benefit from it (from a quick CTRL+F in the manual):

• Recipe 9.1 (and similarly Recipes 9.2, 9.5.2, 9.7, 11.2, 12.3, 17.3):
  From fixed = (sum(sim.substitutions.mutationType == m2) == 1); to fixed = (sim.countOfSubstitutionsOfType(m2) > 0);
• Recipe 9.6.1:
  From if (sum(sim.substitutions.selectionCoeff) == 0.0) to if (sim.sumOfSubstitutionsOfType(m1) == 0.0)

More generally, one could think of a new callback substitute(<mutType>) which takes place every time a mutation of a given type is about to become a substitution. This could be ideal to end simulations as soon as a fixation event takes place, e.g.

substitute(m1) { sim.simulationFinished(); }

(present in many Recipes of sweeps) or, to update a baseline trait-value (that is, to overcome the need for sumOfSubstitutionsOfType(m1) in my example of quantitative traits above, as one could write

substitute(m1) { defineGlobal("BASELINE_TRAIT", BASELINE_TRAIT + mut.selectionCoeff); }

and then

phenos = BASELINE_TRAIT + inds.sumOfMutationsOfType(m1);

or even to have a mutation-based policy of substitution (e.g., by returning T or F from the callback). This idea being shared, I do think a new callback for this might be an overkill.

On a different aspect, but also regarding Substitutions, I wonder if outputSubstitutions() would be more aligned with the SLiMverse vocabulary than the current outputFixedMutations(). But that's a detail :)

[bhaller]

Hi @vsudbrack. I've been working on a redesign of some parts of SLiM to better support quantitative traits. The design will automatically update the baseline offset of the trait after substitution, as you suggest here; so I think that aspect of your proposal is not necessary with that new design. But I like your other suggestions. Not sure whether the new callback type would really pay for itself; each new callback type just makes SLiM even more complicated, of course. It would help with that narrow problem of halting when a substitution occurs, but it doesn't strike me as having a lot of other uses (since the baseline issue will be taken care of). Anyhow, I'll think on all this, thanks!