Mollib 1.0 release notes

Aug 13, 2017

Version 1.0 of mollib is the first release. This page documents the changes and features available in this version.

Minor Releases

  • 1.0.8 - Fixed a bug in assigning an atom’s element when it is not specified.
  • 1.0.7 - Fixed a bug in defining the topology of Ser residues.
  • 1.0.6 - Fixed deployment of manylinux1 and osx releases.

Compatibility

Mollib ver 1.0 is compatible with Python 2.7 and 3.3+.

Requirements

  • numpy
  • scipy

CLI Features

The command-line interface (CLI) includes a series of plugins for processing and analyzing molecular structures.

process

  • The process command plugin is used to read molecular structures and modify them.
  • Currently supports the loading of protein databank (PDB) files as well as the fetching and caching of PDB files from the internet.
  • The hydrogens plugin (--hydrogenate) is included to strip and replace hydrogen atoms in a molecule. The topology of protein molecules are included and the CONECT topologies in PDB files are supported.
  • Individual models or many models can be selected.

measure

  • The measure command plugin is used to measure interatomic distances, angles and dihedrals in molecules.
  • Statistical summaries of measurements are available with the --stats option.
  • Ramachandran angles are calculated with the --rama option. The Ramachandran angles include secondary structure assignments based on the backbone dihedral angles and hydrogen bonding pattern, like DSSP.
  • Statistics on the frequency of the observed Ramachandran angles for the given secondary structure type are provided. Statistics are calculated from the highest resolution structures from the PDB. (Unique structures with resolutions between 0.5-1.6Å‚ observed R-factors better than 0.25 and free R-factors better than 0.30.)

hbonds

  • The hbonds plugin is used to measure and rate amide and aliphatic hydrogen bonds in structures.
  • Hydrogen bonds are classified based on backbone torsion (Ramachandran) angles and the acceptor-donor position in the primary sequence.
  • Statistics on the frequency of observed hydrogen bonds, based on their interatomic distances and angles, are provided. Statistics are calculated from the highest resolution structures from the PDB, where hydrogens were added with the hydrogens plugin. (Unique structures with resolutions between 0.5-1.6Å‚ observed R-factors better than 0.25 and free R-factors better than 0.30.)

pa

  • The pa (partial alignment) plugin is used to fit residual dipolar coupling (RDC) and residual chemical shift anisotropy (RACS) data of biomolecules using a Singular Value Decomposition (SVD).
  • Supports the refinement of sum couplings, like those of methylenes, with the # wildcard operator. ex: ‘CA-HA#’
  • Supports the refinement of multiple molecular structures simultaneously with one dataset.
  • Includes detailed statistics on the fit for each type of interaction.
  • Identifies outliers based on Grubbs tests.
  • Includes ‘fixers’ to find and fix errors in the sign of RDCs and RACS as well as the exclusion of outliers.
  • Reads data from mollib partial alignment files (.pa), NMRPipe DC input files and magnetic resonance data in Xplor-NIH format.
  • Supports the automatic fetching of magnetic resonance data from the PDB. (i.e. .mr files) However, only residue numbers are read in at this time.
  • Currently includes CSA tensor data for the backbone H, C, and N atoms of proteins. Other nuclei can be easily integrated through a settings file.
  • Supports methyl RDC data that have been projected onto C-C (or S-C) bonds, as used in Xplor-NIH.
  • Weights the contribution of RDCs and RACS based on relative errors. If errors are not specified, default errors are used from the settings.
  • Pre-calculated dipolar couplings can be used from the settings, or dipolar couplings can be calculated from the bond lengths and gyromagnetic ratios of the spins from an interaction.
  • Interactions are identified using an abbreviated syntax. For example, 14N-H, A.14N-H, 14N-14H and 14N-C-1 are all valid identifiers.
  • Alignment files with multiple alignment media datasets can be selected with the --set option.

API Features

The API has the following features and updates.

  • Testing Framework. A testing framework that uses pytest, tox and a CLI tester.
  • CLI Tester. The CLI tester is located in the tests/cli directory, and it is invoked by typing make test-cli in the root project directory. It is also included in the tox tests under the cli environment.
  • Plugin Framework. A plugin framework to add and remove functionality. The base Plugin is located in the plugins folder.
  • Settings Manager. A settings manager to dynamically change settings from the command line or through configuration files.
  • Optimized Functions. Optimized (Cython) functions to calculate the length of vectors and to find atoms within a distance cutoff located in mollib.core.
  • Statistics Plugin. A statistics plugin to calculate statistics and datasets over high-resolution structures.