LSalign interpretations with large RMSDs

This forum is shown on the index page along with all topics.

Moderator: robpearc

lam.a
Posts: 1
Joined: Thu Mar 21, 2024 10:43 am

LSalign interpretations with large RMSDs

Post by lam.a »

Hello,

I was wondering if anyone has any experience with using LS-align on similar chemical analogues but had large RMSD values? I have two very similar compounds used on LS-align (differing only by the addition of a Fluorine atom) but the RMSD was ~3 for both the rigid and flexible LS-align options, shouldn't I be expecting an RMSD value closer to 0?

I retrieved these .mol2 files based on SMILES using https://datascience.unm.edu/tomcat/biocomp/convert, with the +2D and +3D boxes checked. I have also tried checking only the +2D option and results were better but still not expected (RMSD ~1)

These were the two SMILES I used:
Fc1ccc(cc1)c1nc([nH]c1c1ccncc1)c1ccc(cc1)S(=O)C
FC1=CC=C(C=C1)C(N=C2C3=CC=C(C(O)=O)C=C3)=C(N2)C4=CC=NC=C4
hujun
Posts: 1
Joined: Sat May 04, 2024 9:41 am

Re: LSalign interpretations with large RMSDs

Post by hujun »

Hi. Many thanks for your question. I am Jun Hu, one of the author of LS-align. I have check the format of the output of the tool of "https://datascience.unm.edu/tomcat/biocomp/convert". The output format is quitly different from OpenBabel, which is the correctly tool for transfering smile or other format to mol2. I have used OpenBabel (downloaded at https://github.com/openbabel/openbabel/releases) to transform your given smiles to 2D mol2 and employed LS-align (Rigid) to superimpose them. The results can be found at https://seq2fun.dcmb.med.umich.edu/LS-a ... 95_695682/. The Detail results we also shown as follows:

Name of Query Ligand: *****
Name of Template Ligand: *****
Atom Number of Query Ligand: 27
Atom Number of Template Ligand: 27

Rigid LS-align Results
PC-score:
PC-score = 0.9583 P-value = 0.000102 (if normalized by atom number of Query Ligand)
PC-score = 0.9583 P-value = 0.000102 (if normalized by atom number of Template Ligand)
(You should use PC-score normalized by the atom number of the reference ligand)
Jaccard Ratio: 1.0000
RMSDLS: 0.0000
Alignment Information:
(":" denotes aligned atom pairs of d < 1.0 A, "." denotes other aligned atoms)
Query Atom Index: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Query Atom Type: F C C C C C C C N C N C C C C N C C C C C C C C
Aligned Tag: : : : : : : : : : : : : : : : : : : : : : : : :
Templ Atom Type: F C C C C C C C N C N C C C C N C C C C C C C C
Templ Atom Index: 1 2 3 4 5 6 7 8 9 10 21 20 22 23 24 25 26 27 11 12 13 14 18 19
Query Atom Index: 25 26 27
Query Atom Type: S O C
Aligned Tag: : : :
Templ Atom Type: C O O
Templ Atom Index: 15 16 17
Post Reply