Colleagues,
I have created a Pipeline Pilot protocol (attached) that will read a MOL2 (attached) or an SD file
of molecules with all hydrogens in a 3D conformation and calculate the complete moment of
inertia tensor (matrix). The final components of the protocol retain the tensor elements - Ixx, Iyy,
and Izz, as well as the normalized quantities - Ixx / Izz, Iyy / Izz, and Izz / Izz.
Sauer and Schwarz have shown that a 2D plot analysis of the results in
Ixx / Izz (X-axis) vs Iyy / Izz (Y-axis) space can be used to elucidate molecular shape differences among
combinatorial compound libraries (Sauer and Schwarz, JCIM 43 (2003) 987, Sauer and Schwarz, Chimia 57 (2003)
276.). My colleagues and I have also used moments of inertia to analyze the molecular shapes of GPCR antagonists,
Ro5 compliant MDDR compounds, proposed combichem libraries, and the Abbott compound collection (Akritopoulou-Zanze, I.,
Metz, J.T., and Djuric, S.W., DDT 12 (200&) 948).
The protocol should run fine under PP SES 7.0 as well as PP version 6.0. The matrix diagonalization code is messy
since I wrote it using PilotScript in version 6.0 by unrolling nested loops needed to perform the Jacobi rotations
(matrix diagonalization). In version 7.0, the PilotScript could (and should) be cleaned up and re-written using
the nested looping capability. I lost some hair trying to unroll do loops to make the Jacobi code work
in version 6.0, but it does work!
Many thanks to the excellent Accelrys support group, especially Anne Li-Zhong, who provided the PERL code to extract the
atomic coordinates and masses needed to calculate the MoI tensor.
Enjoy.
Regards,
Jim Metz
Abbott Laboratories
James.Metz@Abbott.com
(847) 936 - 0441
I have created a Pipeline Pilot protocol (attached) that will read a MOL2 (attached) or an SD file
of molecules with all hydrogens in a 3D conformation and calculate the complete moment of
inertia tensor (matrix). The final components of the protocol retain the tensor elements - Ixx, Iyy,
and Izz, as well as the normalized quantities - Ixx / Izz, Iyy / Izz, and Izz / Izz.
Sauer and Schwarz have shown that a 2D plot analysis of the results in
Ixx / Izz (X-axis) vs Iyy / Izz (Y-axis) space can be used to elucidate molecular shape differences among
combinatorial compound libraries (Sauer and Schwarz, JCIM 43 (2003) 987, Sauer and Schwarz, Chimia 57 (2003)
276.). My colleagues and I have also used moments of inertia to analyze the molecular shapes of GPCR antagonists,
Ro5 compliant MDDR compounds, proposed combichem libraries, and the Abbott compound collection (Akritopoulou-Zanze, I.,
Metz, J.T., and Djuric, S.W., DDT 12 (200&) 948).
The protocol should run fine under PP SES 7.0 as well as PP version 6.0. The matrix diagonalization code is messy
since I wrote it using PilotScript in version 6.0 by unrolling nested loops needed to perform the Jacobi rotations
(matrix diagonalization). In version 7.0, the PilotScript could (and should) be cleaned up and re-written using
the nested looping capability. I lost some hair trying to unroll do loops to make the Jacobi code work
in version 6.0, but it does work!
Many thanks to the excellent Accelrys support group, especially Anne Li-Zhong, who provided the PERL code to extract the
atomic coordinates and masses needed to calculate the MoI tensor.
Enjoy.
Regards,
Jim Metz
Abbott Laboratories
James.Metz@Abbott.com
(847) 936 - 0441
