Edit: After a conversation with a coder this post has been heavily altered.
That would be me. Hi there. In general, these are decent thoughts but I’d suggest taking a look at the code before issuing directives.
On duplicative traits: Several traits in-game duplicate effects of other traits. That’s intentional and with good reason. In some cases, yes, they stack. In more cases, they specifically don’t stack. The major problem is with the category-enforcement system: if a trait is part of multiple categories–such as High Night Vision–having it contributes an equal amount of category-strength to every category that’s involved.
This allows random mutations to cross into those categories, which is perfectly fine in moderation (after all, mutation kinda lends itself to biological Weirdness) but, pre-category-enforcement, led to Crawling Chaos where people would do their damndest to game the system and get All The Positives From Every Category. Lots of purifier and mutagen flipping, save/load where available, and generally a situation that was more tedious than fun.
A more immediate problem for our Ursine folks is that significantly-shared traits also add a lot of threshold-blocking if most of your category’s mutations have this sort of thing. Ursine turns out to be more of a culprit than I’d thought: if you miss the first few shots at the Threshold, you may mutate back out of Ursine as your primary category altogether (!). In general, if a mutation is in more than four or five categories, I’d want to break that up.
So, going back to High Night Vision, as I write this HNV makes you part bear, part fish, part Beast, part insect, part rat, part Chimera, and part wolf. (+8 to URSINE, +48 to Not-URSINE. Yikes.) I’m going to remove Ursine from that (and from Nearsighted) and combine those into one URSINE_EYE mutation; that strengthens Ursine for threshold-breaching purposes, at the cost of (marginally) reducing the crossover prospects.
[hr]
Intolerances will be revisited when I get the next chance to work on digestion. The general ideas seem worth consideration.
Making the tank classes tougher can be considered, though I’d do that with separate mutations. There are OVERSIZE variants of the Survivor-type gear, so HUGE folks can get later-game armor and storage. Powered suits, well, those are for the 225-pound weaklings. 
Truth be told, I’d rather futz around with Weird Redirects on slime HP than give 'em a full rework to make one HP pool. Slime bones may as well be hollow; if there’s no serious (Threshold-blocking as with Ursine, for example) need to break stuff out, I prefer to leave some room for intercategory crossover. (NV mutations were such an issue.) Slimes get great IN/PE/DX, but at the cost of ST and taking significant IN penalties from pain.
If you can afford post-threshold Alpha (if it’s easy to achieve now, that’s a bug. Let me know and it’ll get more expensive) and opt to forgo the Body Stuff of the other categories, then you get good stats. Just don’t lose 'em–they don’t record your previous values. Will look into downsides when I take a shot at its capstone, but that’s not an immediate priority.
Elf-A, well, take a look at the code and the design specs. I think that’s going a bit overboard.
Chimera: the idea of a faster Unstable Genetics is interesting, but as-is it’s designed as a dangerous category pre-thresh, with reason. Chimera was designed to create expendable shock troops: to be dropped into high-concentration zed areas and kill a lot of critters before they die. I’m open to the rapid-changing post-thresh, though.
(I did name the THRESH_CHIMERA trait Chaos, after all. 
)
