Tables of IATYPE, ATYPE/RTYPE in CRYSOL

Calculation of SAXS and SANS profiles (CRYSOL, CRYSON), superposition of models (SUPCOMB, DAMAVER, DAMCLUST), database DARA
Post Reply
Message
Author
HLee
Member
Posts: 11
Joined: 2012.05.04 21:37

Tables of IATYPE, ATYPE/RTYPE in CRYSOL

#1 Post by HLee » 2013.12.04 19:48

About CRYSOL,
Could you let me know where I can find current look-up(?) tables of IATYPE and ATYPE/RTYPE (for protein atoms)?

Currently, I can only find the tables (Tables 1 and 2 in Section 5 - READING PDB FILES) in the crysol.txt file (for CRYSOL ver. 2.7), and the tables seem somewhat outdated(???).
- eg. it seems that "CG/ALA" is recognized as "C" (IATYPE = 1) rather than "CH3" (IATYPE = 4) in the current CRYSOL ver. 2.8.2 [although I may be wrong].

So if there has been any changes in the tables, could you me know?

Regards,

HLee
Member
Posts: 11
Joined: 2012.05.04 21:37

Re: Tables of IATYPE, ATYPE/RTYPE in CRYSOL

#2 Post by HLee » 2013.12.05 17:35

Another question is the effect/role of Residue Sequence Number in CRYSOL (especially for bead models, such as DAMMIN models).

For the following four pdb files (same atoms/coordinates, but changing the order or changing the residue sequence numbers),
default runs of CRYSOL (ver. 2.8.2; crysol test1.pdb, etc) give the same outputs for test1.pdb and test2.pdb but not for test3.pdb & test4.pdb.
CrysolOutput(test1.pdb) = CrysolOutput(test2.pdb) ~= CrysolOutput(test3.pdb) ~= CrysolOutput(test4.pdb)
[cf. ~= stands for "not equal to"]

======================================================== test1.pdb
ATOM 1 CG ALA A 1 -40.946 -16.076 0.200 1.00 20.00 0 2 201
ATOM 2 NZ LYS A 2 -51.082 -3.239 -1.844 1.00 20.00 0 2 201
========================================================

======================================================== test2.pdb
ATOM 1 NZ LYS A 1 -51.082 -3.239 -1.844 1.00 20.00 0 2 201
ATOM 2 CG ALA A 2 -40.946 -16.076 0.200 1.00 20.00 0 2 201
========================================================

======================================================== test3.pdb
ATOM 1 CG ALA A 1 -40.946 -16.076 0.200 1.00 20.00 0 2 201
ATOM 2 NZ LYS A 1 -51.082 -3.239 -1.844 1.00 20.00 0 2 201
========================================================

======================================================== test4.pdb
ATOM 1 NZ LYS A 1 -51.082 -3.239 -1.844 1.00 20.00 0 2 201
ATOM 2 CG ALA A 1 -40.946 -16.076 0.200 1.00 20.00 0 2 201
========================================================

Were the different outputs caused by the Residue Sequence Number?
If yes, how are the residue sequence numbers assigned in the DAMMIN model?
- Is it just fixed to "9 bead atoms / residue number" in DAMMIN?

HLee
Member
Posts: 11
Joined: 2012.05.04 21:37

Re: Tables of IATYPE, ATYPE/RTYPE in CRYSOL

#3 Post by HLee » 2013.12.10 17:33

I have done a simple test to find the actual Atomic Group for each ATYPE/RTYPE in CRYSOL (ver. 2.8.2)
- Test: calculation of the atomic scattering factor in vacuo at q = 0 for each ATYPE/RTYPE using the two-atom model (having the same ATYPE/RTYPE but different coordinates)
=> Calculation of the atomic scattering factor in vacuo at q = 0
f_invacuo(q=0) = sqrt( I_invacuo(0)/4 )
-> the actual Atomic Groups / actual IATYPE are searched based on the calculated f_invacuo(q = 0), which represents total number of valence electrons in the Atomic Group.
=> The test results are attached as a pdf file (Table_proteinAtomsCrysol.pdf).
Table_proteinAtomsCrysol.pdf
(19.05 KiB) Downloaded 232 times
=================================
The test results showed that:
- For several ATYPE/RTYPE, the actual Atomic Groups / actual IATYPE are different to Table 2 in the crysol.txt (CRYSOL ver. 2.7).
=> Most are minor changes (upto +/- 3*H differences).
=> But for the ATYPE: CA (Calpha; for all RTYPE), f_invacuo(q=0) is 12.7783, which does not make sense for me
since f_invacuo(q=0) of CA (Calpha) should be either around 7 or around 8 (for CA/GLY).
--> Was it a program bug? or was it intended to weight about 2-fold (?) for CA (Calpha)?
--> If it is a program bug, could you let me know when I can download the bug-fixed CRYSOL program?
If it is not a program bug, does it (CA -> 12.7783) affect the calculation of scattering profiles of the hybrid model
(full atomic protein model + bead atomic models) where the bead atomic models have electron densities different to the protein.

Regards,

SaxsMax
Administrator
Posts: 349
Joined: 2007.08.03 19:16

Re: Tables of IATYPE, ATYPE/RTYPE in CRYSOL

#4 Post by SaxsMax » 2013.12.11 04:51

1. The table in Crysol containing the number of hydrogens in an atomic group
has been compiled according to the PDB hetero-atom dictionary:
http://rcsb-deposit.rutgers.edu/het_dictionary.txt
http://deposit.pdb.org/het_dictionary.txt

2. It is incorrect to compute the scattering from a bead model by Crysol.
one should use pdb2alm or pdb2is instead

3. Subsequent lines with CA in pdb files are considered as dummy residues,
those differential formfactor would be around 12.7783

Post Reply