So, a thing that 5.5e does: it decouples ability score bonuses from species (previously known as 'race', in PF2 known as 'ancestry'; I think 'species' is a fine choice). I don't necessarily disagree with this decision, and I recognize it was done for good reasons, but there is much nuance to be dug through, so let's explore.
This is a hugely contentious issue, so hopefully I don't make too many dumb assumptions or dumb material conditionals. (I almost certainly will!)
Nomenclature
Ok, so the main reason ability scores are decoupled from race is that we Definitely Do Not want to leave ourselves open to real-world biological race conspiracy theories, like "Black people have lower Int but higher Str than white people". Which is fair!
...but, we already renamed it? Nobody can any longer fairly accuse D&D species of being the same thing as Earth race?
D&D species is of course not the same thing as biology science species, but the name change does make the decoupling of ability scores from species slightly goofier -- who could possibly argue that your average standard-issue hippopotamus is not able to lift more weight (ie, is not stronger) than your average standard-issue groundhog?
Introducing Sex, for the Purpose of Clarifying but Probably in fact Muddying the Issue
In OD&D 1e, women had a cap on how strong they could be. That's goofy in at least two ways (probably three: I think female characters didn't get any advantages to compensate for this disadvantage), so let's cut it down to only one way: Pretend (biological, ie, ovary-having?) women had, say, a -2 to Strength score and a +2 to, I dunno, Dexterity. Or, let's make women the Default and pretend they had no modifiers to abilities, and men had a +2 to Str and a -2 to Dex.
It is a source of perpetual mass dumbness that, on average, men, however you define it, are probably statistically basically stronger than women, however you define it. This is the (ostensible) objection to having transwomen compete in women's sports.
Setting aside the very many dumbnesses associated with this argument, the core thing that's maybe true (depending, again, on how you define "woman" and "man", among other things) is that the bell curve of women's strength scores is shifted a bit lower than the bell curve of men's strength scores. The strongest man is stronger than the strongest woman; the weakest man is stronger than the weakest woman; the average man is stronger than the average woman; BUT it's only shifted a little, so of course the strongest woman is stronger than the weakest man.
If ability scores are generated by rolling, this shifted bell curve is of course represented very well by a flat +2 or -2. (This is why 1e's actual thing, a straight-up cap on Strength, is doubly dumb and bad.)
If we wanted to maximize Realism at the cost of everything else, a -2 penalty compensated by a +2 bonus elsewhere (or, like, giving men a +2 Str and women a +2 Dex, nobody getting penalties and everybody getting equal bonuses to different things) would be a fair model of offset bell curves.
Sex is a super-goofy thing to demand be "realistic"ally modeled, of course, but this is how we'd do it. (Why it's super-goofy: among other reasons, consider species with different sexual dimorphisms. Species where the girl is bigger and stronger than the boy, for example.)
Race, Tho?
Let's consider now if ancestry were still called Race.
Let's pretend there's an ability called Melanin, which reflects how much melanin is in a character's skin.
A character of the Black Race should have a bonus to their Melanin score, yes? And maybe a character of the Nordic Race should have a penalty? The average Black person is darker-skinned than the average Nordic person, yes? Still permitting the existence of the darkest-skinned Nordic person being darker-skinned than the lightest-skinned Black person, etc.
Sure, there are Backgrounds which could conceivably have an effect on the Melanin score -- the Beach Bum background, for example -- but most of them probably won't have as much effect as Race.
On the other extreme, we definitely would not want to say any Race has a bonus or penalty to, say, Intelligence, Charisma, or Wisdom. That would be some caliper-wielding chauvinism, away from which we want to vigorously shy.
But... are there maybe actually some Races that average stronger or faster or more hardy than others?
Ok, let's set Race aside again, let's set it from our minds, we're now again talking about Species only.
Curvature of the Struck Idiophone Variety
Ok, so the bell-curve thing doesn't necessarily so much completely apply, for like three reasons: first, we usually don't roll for ability scores anymore, as array and point by are objectively better; second, PCs are Special, so they shouldn't necessarily reflect the broader bell curve of the species population; third, bell curves in terms of populations have, like, a bad reputation, mostly because of a book of that name.
Still... does it actually make sense that the strongest Orc it is possible to produce is exactly as strong as the strongest Gnome it is possible to produce? The most dexterous Dwarf as dexterous as the most dexterous Elf?
Proposed House Rule
My inclination, to make Species at least have the ability to matter, though it doesn't necessarily have to for every character, is a house rule along these lines:
Assign each Species two (or so) ability scores, which will typically be whichever ones they got bonuses to in 5.0e. When you assign your Origin-related ability scores (your choice of { +2 +1 } or { +1 +1 +1 }, as usual), you may pick from the options given in your Background and the options given in your Species. You have the option for only your Background to matter to your ability scores, the option for only Species to matter, or the option for both to matter.
You still have a situation where any Orc and a Gnome Soldier have the same maximum possible Strength, which remains slightly peculiar -- but better than unaltered 5.5e, where a Gnome Soldier had a higher maximum Strength than an Orc Criminal.
Sub-Houserule: Synergy
If there is overlap between your Species and your Background, my first instinct is nothing special happens, but it does happen to provide, like, a soft limit: you have fewer options to choose from, so you are pushed towards choosing what is overlapped, though it is not in any special way required.
My second instinct is to give overlap between Species and Background some sort of synergy bonus. Which is kind of like 3.5e's Racial Favored Class, except it's more of a Species Favored Background, and also not something that every table just ignores because it's too much hassle to bother with.
My first thought here was to give ability scores with Background+Species synergy a starting max in that Ability of 22 instead of 20 -- then I remembered no character will ever start with higher than 15+2=17 in any ability, and I don't want Origin to add a thing you can first plausibly take advantage of at level 8.
Perhaps, if you have Background+Species synergy, then the option is made available to you to, instead of { +1 +1 +1 } or { +2 +1 }, choose { +3 } provided it is used on a synergized ability, thereby allowing a character plausibly to begin with a score of 18 in that ability?
Or, like, if you have synergy, then you have the option to forgo your Origin Feat and 1 point of your ability bonus (leaving you with a choice between only { +2 } or { +1 +1 }), and instead take a regular feat, provided it is a feat that gives you a bonus to the synergized ability (and, of your remaining ability points, the +2 or the +1s can go to any abilities, even the synergized one)?