On Sep 8, 2026, at 11:26 AM, Wes Dennis via ChimeraX-users <chimerax-users@cgl.ucsf.edu> wrote:

Hello,

Hi Wes,

We're using ChimeraX for structural analysis of the 23S rRNA peptidyl transferase centre in 1VY4 (2.60 Å, two 70S copies). We ran addh on the full deposition and found that thirteen base amino and imino nitrogens received no hydrogen. All thirteen sit 2.68–3.54 Å from a modelled magnesium, so metalDist is the obvious explanation, and we've since restored those protons from chemistry templates for downstream work.
We're now trying to understand precisely what addh did rather than infer it, and four questions would help. Not reporting a bug — the behaviour looks correct, we just can't tell from the documentation which of several readings is right.
1. Is there any way to obtain the list of hydrogens addh declined to add under metalDist, together with the X–H–M angle it computed for each? A log option, a return value, an attribute. Failing that, is the suppression recorded anywhere at all — even as a count?

It is not currently tracked in any way. It seems reasonable that addh should issue log messages for hydrogens skipped due to metal proximity. After all, it already logs a message when it doesn't add hydrogens to an atom because that atom is missing heavy-atom bond partners. I will open an enhancement-request ticket in our bug tracking database with you cc'ed so that you will be notified when it gets implemented. It might take a little while because a lot of work piled up while I was away at the workshop.

2. For which functional groups does hbond true actually adjust hydrogen position? Specifically, is an aromatic exocyclic amino — adenine N6, cytosine N4, guanine N2 — treated as rotatable about the C–N bond, or are its two hydrogens fixed by the heavy-atom geometry? If adjustable, is the adjustment confined to the base plane or can it go out of plane? The hbonds documentation describes well-determined groups versus rotatable hydroxyls, and we're unsure which side an aromatic amino falls on — and whether that page's criteria apply to addh's hbond option at all.

Well, obviously hydrogen positions get adjusted for hydroxyls and what not. For the NH2 group of an aniline moiety, addh will try to maximize the H-bonding interaction of the protons and/or lone pair, which can cause the NH2 to be out of the plane of the ring. It's not really rotating around the C-N bond per se.

3. The documentation says a hydrogen is suppressed if the X–H–M angle would be >120°. At what position is that angle evaluated — the position addh would have chosen with hbond optimisation applied, or a geometry-only position before optimisation?

The position addh would have chosen.

4. For an exocyclic amino group, how are H61/H62, H41/H42 and H21/H22 assigned to the two hydrogen positions — deterministically from the 3D coordinates, or by a fixed convention relative to the ring? Related: is metalDist evaluated per candidate hydrogen? We see nitrogens where one hydrogen was added and the other was not, which the documentation doesn't address, and if so we'd like to know whether the surviving hydrogen keeps a stable name.

I don't believe the names are assigned in any deterministic way. There will be an Hn1/2, but either proton might get the 1 and the other the 2. If a hydrogen is skipped, its name is skipped too, so the surviving proton may be 1 or 2.