Confusing trivalent sulfur that shouldn't get a generic but does

[bannanc]

I initially thought this molecule was only in the unfiltered DrugBank, but we do have a Tripos mol2 file for it (no gaff2 mol2 file). The highlighted torsion is assigned the generic, which doesn't make sense to me as there is a SMIRKS pattern in the smirffishfrcmod file that matches it.

This image is a little confusing because in 2D there are 2 oxygens overlapping only the black one is bound to the sulfur. The highlight torsion should match this line from smirffishFrcmod:
[*:1]~[#16X4,#16X3+0:2]-[#6X4:3]-[*:4] 1 0.000 0.0 3 Frosst X_-SO-CT-X guess
The highlighted torsion does match the smirks pattern if I load it into the substrucDepict notebook so I'm unsure why it is being assigned incorrectly.

[bannanc]

Here is the same torsion as it matches the smirks patter "[*:1]~[#16X4,#16X3+0:2]-[#6X4:3]-[*:4]" in the substrucDepict notebook

[bannanc]

Here's another example molecule:

[bannanc]

And again, the smirks patter "[*:1]~[#16X4,#16X3+0:2]-[#6X4:3]-[*:4]" matches the generic torsion in a separate search...

@davidlmobley
Can you think of any reason the SMIRKS search would match a molecule, but it wouldn't get assigned that parameter?

[bannanc]

When the same molecule as above gets charged, it becomes a sulfur with a negative charge:

Is it possible the forcefield module is interpreting it as having a -1 charge to begin with?
Either way, that sulfur was the same in the original parm@Frosst SO was for [#16X4,#16X3], but in smirff Chris wrote a lot of them as [#16X4,#16X3+0] for some reason.

I'm going to try removing the +0 from that torsion and see if anything changes.

[bannanc]

It seems that all of them with this set in the tripos mol2 files don't get a generic if I fix the SMIRKS so it is

[*:1]~[#16X4,#16X3!+1:2]-[#6X4:3]-[*:4] 1 0.000 0.0 3 Frosst X_-SO-CT-X guess

This is the same fix we used for the relevant bonds, as in smarty issue#216

[bannanc]

Added #16X3!+1 to pull request #43
Also putting this type of chemistry in the "needs exploration" list