Error when generating fragment libraries

mostafa.razavi
Topic Author
Offline
Junior Member

5 years 5 months ago #550 by mostafa.razavi

Dear Cassandra community,

I'm trying to run NVT simulation of a rather complicated molecule that has 5 rings. The molecule is 1-naphthalenyl,4-phenanthrenyl butane. I'm attaching a drawing of the molecule with numbered united-atom sites as well a zip file containing ff, mcf, pdb, and input file. When I run the following command to generate the fragment libraries. I get an error. Can anyone help me figure out what I'm doing wrong? Also, I noticed that some atoms (number 3, 17, and 18) in Atom_Info section of mcf file have two "rings", while other ring atoms have only one "rings". Is it because they are part of two rings AND they are the center of the fragment they belong to?

$ python2.7 library_setup.py cassandra_gfortran.exe nvt.inp 1-p-4-n-B.pdb

********************** Cassandra Setup *********************

Cassandra location: /home/mostafa/Git/Cassandra_V1.2/Src/cassandra_gfortran.exe
Scanning input file
  Number of species found: 1
  The MCF file number 1 is: 1-p-4-n-B.mcf
    Species 1 has 9 fragments
Molecules with rings found. These are:
Species 1 has 5 rings.


Creating input MCF generation file for species 1 
Running Cassandra to generate MCF files
One of the local ids is zero
This occurred while writing the angle info section
for the ring fragment mcf file
Fragment 2
Species 1
This error occurred in subroutine Write_Ring_Fragment_MCF_Angle_Info on step 0.
Fatal Error. Stopping program.
Traceback (most recent call last):
  File "/home/mostafa/Git/Cassandra_V1.2/Scripts/Frag_Library_Setup/library_setup.py", line 619, in <module>
    angle_type = line.split()[4]

File Attachment:

File Name: 1-p-4-n-B_...15-2.zip
File Size:4 KB

Thank you,

Mostafa Razavi

Attachments:

Please Log in to join the conversation.

jshah46545
Offline
Junior Member

5 years 5 months ago #551 by jshah46545

Mostafa,

As the script for generating fragment does not currently support fused rings, it incorrectly identifies that the molecule contains 9 fragments. The molecule has the following fragments (any (fused) ring is considered as a fragment along with its exoring atoms) :

Fragment 1: atoms 1-11
Fragment 2: atoms 1, 11, 12
Fragment 3: 11, 12, 13
Fragment 4: 12, 13, 14
Fragment 5: atoms 14-28

If you manually modify the # Fragment_Info and # Fragment_Connectivity sections in the mcf file and rerun the fragment generation script, it might work.

Please let us know if you continue to have issues.

Best,
Jindal

Please Log in to join the conversation.

mostafa.razavi
Topic Author
Offline
Junior Member

5 years 5 months ago #552 by mostafa.razavi

Jindal,

Thank you for your helpful response. I manually specified Fragment_Info and Fragment_Connectivity sections in mcf file as below:

!index number_of_atoms_in_fragment branch_point other_atoms
# Fragment_Info
5
1 11 1 2 3 4 5 6 7 8 9 10 11
2 3 11 1 12
3 3 12 11 13
4 3 13 12 14
5 15 15 16 17 18 19 20 21 22 23 24 25 26 27 28 14

# Fragment_Connectivity
4
1 1 2
2 2 3
3 3 4
4 4 5

The script ran successfully, but Cassandra gave an error:

Number of atoms in species 1 is 28
Number of atoms deleted with frag 1 is 9
Number of atoms deleted with frag 2 is 4
This error occurred in subroutine Get_Fragment_Connectivity_Info on step 0.
Fatal Error. Stopping program.

Then I tried considering 13, 14, 15 as a fragment, with the same logic as we consider 1, 11, 12 a fragment. The script and Cassandra both run successfully:

!index number_of_atoms_in_fragment branch_point other_atoms
# Fragment_Info
6
1 11 1 2 3 4 5 6 7 8 9 10 11
2 3 11 1 12
3 3 12 11 13
4 3 13 12 14
5 3 14 13 15
6 15 15 16 17 18 19 20 21 22 23 24 25 26 27 28 14

# Fragment_Connectivity
5
1 1 2
2 2 3
3 3 4
4 4 5
5 5 6

Is the above specification correct? Specifically, have I identified the branch_point values for the first and last fragments correctly?

Please Log in to join the conversation.

mostafa.razavi
Topic Author
Offline
Junior Member

5 years 5 months ago #553 by mostafa.razavi

Also, I realized that in the mcf file generated by mcfgen.py script, atoms 3, 17, and 18 have "ring ring" in the Atom_Info section. I cannot think of a special trait that these atoms might have, other than being part of two rings. If this is true, should 8, 23, and 26 also have "ring ring" instead of "ring"?

# Atom_Info
28
1     CA      Ca   12.01070  0.0  LJ    21.000     3.880  ring
2     CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
3     C1A     1A   12.01070  0.0  LJ    30.000     3.700  ring  ring
4     CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
5     CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
6     CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
7     CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
8     C1A     1A   12.01070  0.0  LJ    30.000     3.700  ring
9     CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
10    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
11    CH2     H2   14.02660  0.0  LJ    46.000     3.950
12    CH2     H2   14.02660  0.0  LJ    46.000     3.950
13    CH2     H2   14.02660  0.0  LJ    46.000     3.950
14    CH2     H2   14.02660  0.0  LJ    46.000     3.950
15    CA      Ca   12.01070  0.0  LJ    21.000     3.880  ring
16    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
17    C1A     1A   12.01070  0.0  LJ    30.000     3.700  ring  ring
18    C1A     1A   12.01070  0.0  LJ    30.000     3.700  ring  ring
19    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
20    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
21    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
22    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
23    C1A     1A   12.01070  0.0  LJ    30.000     3.700  ring
24    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
25    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
26    C1A     1A   12.01070  0.0  LJ    30.000     3.700  ring
27    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring
28    CH      Ch   13.01864  0.0  LJ    50.500     3.695  ring

Please Log in to join the conversation.

ryangmullen
Offline
Administrator

5 years 5 months ago #554 by ryangmullen

Hi Mostafa,

Yes, atoms 13, 14, 15 should be identified as their own fragment and everything looks correct in your # Fragment_Info and # Fragment_Connectivity sections.

I think "ring ring" and "ring" will both be read the same by Cassandra. Are you treating fragments 1 and 6 (the fused ring fragments) as flexible or rigid?

Ryan

Please Log in to join the conversation.

mostafa.razavi
Topic Author
Offline
Junior Member

5 years 5 months ago #555 by mostafa.razavi

Thanks Ryan,

I'm simulating this molecule using TraPPE-UA force field, so yes I'm trying to treat the fused rings as rigid. I set all the bonds, and angles to be fixed, but I have dihedrals as you can see in the mcf file.

Please Log in to join the conversation.

Time to create page: 0.185 seconds