<See way down inside...>
On 2/1/2017 4:58 AM, Paul Emsley wrote:
> On 31/01/2017 20:09, Dale Tronrud wrote:
>> On 1/31/2017 11:51 AM, Paul Emsley wrote:
>>> On 31/01/17 17:54, Edwin Pozharski wrote:
>>>> Whatever the rationale was, there is a structure in the PDB that has
>>>> alternate conformer of a residue listed with different residue type -
>>>> A is arginine and B is glutamine. Coot fails to load the model
>>>> complaining in the command window
>>>>
>>>> WARNING:: Error reading small-molecule cif "/home/epo/coot/foo.pdb"
>>>> There was an error reading /home/epo/coot/foo.pdb.
>>>> ERROR 42 READ: Duplicate sequence number and insertion code.
>>>> LINE #1571
>>>> ATOM 1666 N BGLN B 93 24.448 28.340 -33.325 0.50
>>>> 9.34 N
>>>>
>>>> No Spacegroup found for this PDB file
>>>> There was a coordinates read error
>>>>
>>>
>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1503&L=COOT&F=&S=&P=25056
>>>
>>>
>>
>> I think this is a poor solution.
>
> We agree, I think, that it is not a good solution.
>
>> Microheterogeneity is not a duplicate residue number.
>
> Isn't it? My understanding is that using the same residue number for a
> different residue type is *is* how microheterogeneity is specified.
>
>> Not any more so than the alternative
>> conformation that is also indicated with "alt loc" letters.
>
> I don't follow this, but if you mean that to describe microheterogeneity
> by using an altloc is the Wrong Way, then I agree with you.
I don't consider this a "duplicate residue number" because there is
only one residue being described. At one time the PDB renamed
inhibitors bound to a protein to residue "1" despite the first residue
of the protein also being named "1". Now that's a duplicate and it's
very confusing!
With microheterogeneity there is only a single "slot" in the sequence
that is being described by the residue number. It just happens that in
some of the unit cells that slot is occupied by, for example, an Ala
while the rest of the unit cells contain Ser. This situation is not
much different than when the residue is always of type Ser but in some
unit cells it is g+ and the rest g- -- You have a single location in the
molecule with mutually exclusive models whose occupancies cannot sum to
more than one. The "alt loc" indicators can be used to specify either
situation, and that is how they are used in the PDB.
Here is an example from a Crambin model (1JXT). Note that there are
three "alt loc"s, two rotomers of Ser and one of Pro. I do not consider
this a "duplicate residue number" since it is a description of just one
place in the sequence of the protein.
ATOM 378 N BSER A 22 4.886 12.647 -3.137 0.25 2.69
N
ATOM 379 N CSER A 22 4.886 12.647 -3.137 0.20 2.69
N
ATOM 380 CA BSER A 22 6.014 13.445 -2.619 0.25 2.65
C
ATOM 381 CA CSER A 22 6.014 13.445 -2.619 0.20 2.65
C
ATOM 382 C BSER A 22 6.335 13.134 -1.171 0.25 2.69
C
ATOM 383 C CSER A 22 6.335 13.134 -1.171 0.20 2.69
C
ATOM 384 O BSER A 22 5.447 12.947 -0.321 0.25 3.16
O
ATOM 385 O CSER A 22 5.447 12.947 -0.321 0.20 3.16
O
ATOM 386 CB BSER A 22 5.771 14.977 -2.622 0.25 2.42
C
ATOM 387 CB CSER A 22 5.303 14.879 -2.638 0.20 1.48
C
ATOM 388 OG BSER A 22 4.801 15.169 -1.599 0.25 3.52
O
ATOM 389 OG CSER A 22 6.657 15.597 -2.270 0.20 4.90
O
ATOM 390 HB2BSER A 22 6.591 15.476 -2.342 0.25 9.94
H
ATOM 391 HB2CSER A 22 5.428 15.332 -3.776 0.20 4.73
H
ATOM 392 HB3BSER A 22 5.454 15.386 -3.472 0.25 5.08
H
ATOM 393 HB3CSER A 22 4.730 15.085 -2.351 0.20 1.05
H
ATOM 394 N APRO A 22 4.886 12.647 -3.137 0.55 2.69
N
ATOM 395 CA APRO A 22 6.014 13.445 -2.619 0.55 2.65
C
ATOM 396 C APRO A 22 6.335 13.134 -1.171 0.55 2.69
C
ATOM 397 O APRO A 22 5.447 12.947 -0.321 0.55 3.16
O
ATOM 398 CB APRO A 22 5.553 14.873 -2.888 0.55 3.49
C
ATOM 399 CG APRO A 22 4.590 14.806 -4.078 0.55 3.51
C
ATOM 400 CD APRO A 22 3.870 13.470 -3.919 0.55 2.50
C
ATOM 401 HA APRO A 22 6.812 13.232 -3.136 0.55 1.43
H
ATOM 402 HB2APRO A 22 5.167 15.299 -2.198 0.55 3.89
H
ATOM 403 HB3APRO A 22 6.300 15.439 -3.105 0.55 13.74
H
ATOM 404 HG2APRO A 22 4.045 15.436 -4.032 0.55 3.61
H
ATOM 405 HG3APRO A 22 5.063 14.827 -4.928 0.55 3.64
H
ATOM 406 HD2APRO A 22 3.057 13.566 -3.405 0.55 5.81
H
ATOM 407 HD3APRO A 22 3.657 12.978 -4.727 0.55 8.33
H
Dale Tronrud
>
>> Both come up quite often in the PDB,
>
> One man's quite often is another man's very rarely.
>
>> and microheterogeneity probably should be put
>> in models more often than it currently is.
>
> I don't doubt that you are right.
>
>> Many modelers simply don't
>> realize it is a possibility.
>
> I agree.
>
>> Your users rarely going to know about the
>> need to put this option into their startup file.
>
> Indeed, if Ed has to ask the list, then I need to reconsider how I
> arrange this problem/work-around. (Maybe mmdb2 (or coot?) no longer has
> the problem with atom selection in models that have duplicate sequence
> numbers).
>
> Paul.
>
