Copyright © 1990-2004 by Jay William Ponder

All Rights Reserved

User's Guide Cover Illustration by Jay Nelson

Courtesy of Prof. R. T. Paine, Univ. of New Mexico

TINKER IS PROVIDED "AS IS" AND WITHOUT ANY WARRANTY EXPRESS OR IMPLIED. THE USER ASSUMES ALL RISKS OF USING THIS SOFTWARE. THERE IS NO CLAIM OF THE MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

YOU MAY MAKE COPIES OF TINKER FOR YOUR OWN USE, AND MODIFY THOSE COPIES. YOU MAY NOT DISTRIBUTE ANY MODIFIED SOURCE CODE OR DOCUMENTATION TO USERS AT ANY SITE OTHER THAN YOUR OWN. PLEASE SIGN AND RETURN THE TINKER LICENSE AGREEMENT IF YOU MAKE USE OF THIS SOFTWARE.

v4.2 6/04

Table of Contents Page

Introduction to the TINKER Package 5

Installing TINKER on your Computer 7

Types of Input & Output Files 10

Potential Energy Programs 13

Additional Utility Programs 19

Force Field Parameter Sets 23

Use of the Keyword Control File 31

Notes on Special Features & Methods 63

Descriptions of TINKER Routines 69

Contents of Common Block Variables 136

Index of Function & Subroutine Calls 163

Examples using the TINKER Package 189

Benchmark Results 191

Collaborators & Acknowledgments 195

References & Suggested Reading 197

Welcome to the TINKER molecular modeling package! TINKER is designed to be an easily used and flexible system of programs and routines for molecular mechanics and dynamics as well as other energy-based and structural manipulation calculations. It is intended to be modular enough to enable development of new computational methods and efficient enough to meet most production calculation needs. Rather than incorporating all the functionality in one monolithic program, TINKER provides a set of relatively small programs that interoperate to perform complex computations. New programs can be easily added by modelers with only limited programming experience. The series of major programs included in the distribution system perform the following core tasks:

(1) building protein and nucleic acid models from sequence

(2) energy minimization and structural optimization

(3) analysis of energy distribution within a structure

(4) molecular dynamics and stochastic dynamics

(5) simulated annealing with a choice of cooling schedules

(6) normal modes and vibrational frequencies

(7) conformational search and global optimization

(8) transition state location and conformational pathways

(9) fitting of energy parameters to crystal data

(10) distance geometry with pairwise metrization

(11) molecular volumes and surface areas

(12) free energy changes for structural mutations

(13) advanced algorithms based on potential smoothing

Many of the various energy minimization and molecular dynamics computations can be performed on full or partial structures, over Cartesian, internal or rigid body coordinates, and including a variety of boundary conditions and crystal cell types. Other programs are available to generate timing data and allow checking of potential function derivatives for coding errors. Special features are available to facilitate input and output of protein and nucleic acid structures. However, the basic core routines have no knowledge of biopolymer structure and can be used for general molecular systems.

Due to its emphasis on ease of modification, TINKER differs from many other currently available molecular modeling packages in that the user is expected to be willing to write simple ``front-end'' programs and make some alterations at the source code level. The main programs provided should be considered as templates for the users to change according to their wishes. All subroutines are internally documented and structured programming practices are adhered to throughout. The result, it is hoped, will be a calculational system which can be tailored to local needs and desires.

The core TINKER system consists of nearly 135,000 lines of source written entirely in a portable Fortran77 superset. Use is made of only some very common extensions that aid in writing highly structured code. The current version of the package has been ported to a wide range of computers with no or extremely minimal changes. Tested systems include: Red Hat Linux, Microsoft Windows 9X/NT/2000/XP, Apple OS9 and OSX, HP/Compaq/DEC Alphas under Tru64 Unix and OpenVMS, Hewlett-Packard, IBM, Silicon Graphics and Sun workstations under each vendor's Unix. At present, our new code is written on various Linux platforms, and occasionally tested for compatibility on various of the other machine and OS combinations listed above. At present, we are in the process of converting our primary development efforts from Fortran77 to a more modern Fortran dialect. A machine-translated C version of TINKER is currently available, and a hand-translated optimized C version of a previous TINKER release is available for inspection. Conversion to C or C++ is under consideration, but not being actively pursued at this time.

The basic design of the energy function engine used by the TINKER system allows usage of several different parameter sets. At present we are distributing parameters that implement AMBER ff94 and ff96, CHARMM19 and 27, MM2, MM3, OPLS-UA, OPLS-AA, Liam Dang's polarizable potentials, and our own AMOEBA (Atomic Multipole Optimized Energetics for Biomolecular Applications) parameters. In most cases, the source code separates the geometric manipulations needed for energy derivatives from the actual form of the energy function itself. Several other literature parameter sets are being considered for possible future development (ENCAD, MMFF-94, MM4, UFF, etc.), and many of the alternative potential function forms reported in the literature can be implemented directly or after minor code changes.

Much of the software in the TINKER package has been heavily used and well tested, but some modules are still in a fairly early stage of development. Further work on the TINKER system is planned in three main areas: (1) extension and improvement of the potential energy parameters including additional parameterization and testing of our polarizable multipole AMOEBA force field, (2) coding of new computational algorithms including additional methods for free energy determination, torsional Monte Carlo and molecular dynamics sampling, advanced methods for long range interactions, better transition state location, and further application of the potential smoothing paradigm, and (3) further development of Force Field Explorer, a Java-based GUI front-end to the TINKER programs that provides for calculation setup, launch and control as well as basic molecular visualization.

Questions and comments regarding the TINKER package, including suggestions for improvements and changes should be made to the author:

Professor Jay William Ponder

Biochemistry & Molecular Biophysics, Box 8231

Washington University School of Medicine

660 South Euclid Avenue

Saint Louis, MO 63110 U.S.A.

office: Center for Computational Biology, Room 208

phone: (314) 362-4195

fax: (314) 362-7183

email: ponder@dasher.wustl.edu

In addition, an Internet web site containing an online version of this User's Guide, the most recent distribution version of the full TINKER package and other useful information can be found at http://dasher.wustl.edu/tinker, the Home Page for the TINKER Molecular Modeling Package.

The TINKER package is distributed on the Internet via either the web site or the anonymous ftp account on dasher.wustl.edu with an IP number of 128.252.208.48. This node is a web and file server located in the Ponder lab at Washington University School of Medicine. The package is available via the web and standard browsers from the TINKER home page at http://dasher.wustl.edu/tinker/. Alternatively TINKER can be downloaded by logging into dasher.wustl.edu via anonymous ftp (Username: anonymous, Password: "your email address") and downloading the software from the /pub/tinker subdirectory. The complete TINKER distributions as well as individual files can be downloaded from this site.

The easiest way to get TINKER running on your machine is to use the self-extracting installation kit for either Linux, Windows, or Macintosh OS X 10.3. The installer will guide you through complete setup of TINKER and the Force Field Explorer (FFE) GUI, and perform all required configuration chores. The installer kits for the three supported systems are tinker4.2-linux.sh, tinker4.2-windows.exe and tinker4.2-macosx.sit. The Linux and Windows kits each contain a private copy of a Java and Java3D run-time environment for use with the package. The Macintosh version requires an OS X 10.3 (Panther) system for installation. The native Java implementation is used on Macs, and the Java3D package must be downloaded from Apple and installed prior to using TINKER with Force Field Explorer.

The TINKER package is also available as compressed Unix tar archives, Windows zip files, and as a complete set of uncompressed source and data files. Binaries are provided for machines running Windows 9X/ME/NT/2000/XP, Linux, and Apple Mac OS X. All of these executables are present in standard compressed formats as individual programs or as complete sets of executables. It is expected that other Unix users and PC users who need specially customized versions, will build binaries for their specific system. Sites with access to the Unix tar, compress and uncompress commands should simply obtain the archive file tinker.tar.Z. Alternatively, tinker.tar.gz and tinker.zip containing identical distributions compressed to GNU gzip and Windows ZIP format are also provided. If you choose to download individual files, you will need at a minimum the contents of the /doc, /source and /params subdirectories. Also required are the compile/build scripts from the subdirectory named for your machine type. Other areas contain test cases and examples, benchmark results, machine-translated C code, and the Force Field Explorer Java GUI for TINKER. The entire TINKER package, after building the executables, will require from about 40 to over 150 megabytes of disk space depending on the components installed and the use of shared libraries in the executables.

The documentation for the TINKER programs, including the guide you are currently reading, is located in the /pub/tinker/doc subdirectory. The documentation was prepared using the Applixware Words and Graphics programs. Portable versions of the documentation are provided as ascii text in .txt files and in Adobe Acrobat .pdf file formats. Please read and return by mail the TINKER license. In particular, we note that TINKER is not ``Open Source'' as users are prohibited from redistribution of original or modified TINKER source code or binaries to other parties. While our intent is to distribute the TINKER code to anyone who wants it, the Ponder Lab would like to remain the sole distribution site and keep track of researchers using the package. The returned license forms also help us justify further development of TINKER. When new modules and capabilities become available, and when the almost inevitable bugs are uncovered, we will attempt to notify those who have returned a license form. Finally, we remind you that this software is copyrighted, and ask that it not be redistributed in any form.

The compilation and building of the TINKER executables should be easy for most of the common workstation and PC class computers. We provide in the /make area a Unix-style Makefile that with some modification can be used to build TINKER on most Unix machines. As a simpler alternative to Makefiles for the Unix versions, we also provide machine-specific directories with three separate shell scripts to compile the source, build an object library, and link binary executables. Three similar command files are provided for Windows, Macintosh and Open VMS systems. Compilation on Unix workstations should use the vendor supplied Fortran compiler, if available. The public domain GNU g77 Fortran compiler available from http://gcc.gnu.org/ is also capable of building TINKER on Linux and other Unix-based machines. The Linux executables we provide are built with the Intel Fortran for Linux 8.0 compiler. The Portland Group (PGI) and Absoft ProFortran compilers have also been tested under Linux, both of which generate executables roughly comparable in speed to the Intel compiler. On Linux, the g77 executables tend to exhibit degraded performance compared with executables from commercial compilers. Some benchmark results are provided in a later section of this User's Guide For the Macintosh we distribute executables built under Apple OS X 10.3 with the GNU g77 compiler. TINKER also builds on the Macintosh using the Absoft ProFortran compiler. For PCs running Windows 9X/NT/2000/XP, the distributed TINKER executables are built under the Intel Fortran for Windows 8.0 compiler. Alternative Windows compilers such as Compaq Visual Fortran, Lahey/Fujitsu and The Portland Group compilers, and GNU g77 under Cygwin have been tested and shown to build TINKER correctly. Please see the README files in each of the machine-specific areas for further information.

The first step in building TINKER using the script files is to run the appropriate compile.make script for your operating system and compiler version. Next you must use a library.make script to create an archive of object code modules. Finally, run a link.make script to produce the complete set of TINKER executables. The executables can be renamed and moved to wherever you like by editing and running the ``rename'' script. These steps will produce executables that can run from the command line, but without the capability to interact with the FFE GUI. Building FFE-enabled TINKER executables involves replacing the sockets.f source file with sockets.c, and included the object from the C code in the TINKER object library. Then executables must be linked against Java libraries in addition to the usual resources. Sample compgui.make and linkgui.make scripts are provided for systems capable of building GUI-enabled executables.

Regardless of your target machine, only a few small pieces of code can possibly require attention prior to building. The first two are the system dependent time and date routines found in clock.f and calendar.f respectively. Next is the openend.f routine that facilitates appending data to the end of an existing disk file. Please uncomment the sections of these routines needed for your computer type. Version of these system dependent routines suitable for each system are also provided in the directory for each machine/OS type. The final set of possible source alterations are to the master array dimensions found in the include file sizes.i. The most basic limit is on the number of atoms allowed, ``maxatm''. This parameter can be set to 10000 or more on most workstations. Personal computers with minimal memory may need a lower limit, perhaps 1000 atoms, depending on available memory, swap space and other resources. A description of the other parameter values is contained in the header of the file. Note that in order to keep the code completely transparent, TINKER does not implement any sort of dynamic memory allocation or heap data structure. This requires that sizes.i dimensioning values be set at least as large as the biggest problem you intend to run. Obviously, you should not set the array sizes to unnecessarily large values, since this can tax your compute resources and may result in performance degradation or overt failure of the executables.

Specific questions about the building or use of the TINKER package should be directed to tinker@dasher.wustl.edu. TINKER related questions or comments of more general interest can be sent to the Computational Chemistry List (http://www.ccl.net/) run by Jan Labanowski at the University of Notre Dame. The TINKER developers monitor this list and will respond to the list or the individual poster as appropriate.

This section describes the basic file types used by the TINKER package. Let's say you wish to perform a calculation on a particular small organic molecule. Assume that the file name chosen for our input and output files is sample. Then all of the TINKER files will reside on the computer under the name sample.xxx where .xxx is any of the several extension types to be described below.

SAMPLE.XYZ

The .xyz file is the basic TINKER Cartesian coordinates file type. It contains a title line followed by one line for each atom in the structure. Each line contains: the sequential number within the structure, an atomic symbol or name, X-, Y-, and Z-coordinates, the force field atom type number of the atom, and a list of the atoms connected to the current atom. Except for programs whose basic operation is in torsional space, all TINKER calculations are done from some version of the .xyz format.

SAMPLE.INT

The .int file contains an internal coordinates representation of the molecular structure. It consists of a title line followed by one line for each atom in the structure. Each line contains: the sequential number within the structure, an atomic symbol or name, the force field atom type number of the atom, and internal coordinates in the usual Z-matrix format. For each atom the internal coordinates consist of a distance to some previously defined atom, and either two bond angles or a bond angle and a dihedral angle to previous atoms. The length, angle and dihedral definitions do not have to represent real bonded interactions. Following the last atom definition are two optional blank line separated sets of atom number pairs. The first list contains pairs of atoms that are covalently bonded, but whose bond length was not used as part of the atom definitions. These pairs are typically used to close ring structures. The second list contains ``bonds'' that are to be broken, i.e., pairs of atoms that are not covalently bonded, but which were used to define a distance in the atom definitions.

SAMPLE.KEY

The keyword parameter file always has the extension .key and is optionally present during TINKER calculations. It contains values for any of a wide variety of switches and parameters that are used to change the course of the computation from the default. The detailed contents of this file is explained in a latter section of this User's Guide. If a molecular system specific keyfile, in this case sample.key, is not present, the the TINKER program will look in the same directory for a generic file named tinker.key.

SAMPLE.DYN

The .dyn file contains values needed to restart a molecular or stochastic dynamics computation. It stores the current position, current velocity and current and previous accelerations for each atom, as well as the size and shape of any periodic box or crystal unit cell. This information can be used to start a new dynamics run from the final state of a previous run. Upon startup, the dynamics programs always check for the presence of a .dyn file and make use of it whenever possible. The .dyn file is updated concurrent with the saving of a new dynamics trajectory snapshot.

SAMPLE.END

The .end file type provides a mechanism to gracefully stop a running TINKER calculation. At appropriate checkpoints during a calculation, TINKER will test for the presence of a sample.end file, and if found will terminate the calculation after updating the output. The .end file can be created at any time during a computation, and will be detected when the next checkpoint is reached. The file may be of zero size, and its contents are unimportant. In the current version of TINKER, the .end mechanism is only available within dynamics-based programs.

SAMPLE.001, SAMPLE.002, ....

Several types of computations produce files containing a three or more digit extension (.001 as shown; or .002, .137, .5678, etc.). These are referred to as cycle files, and are used to store various types of output structures. The cycle files from a given computation are identical in internal structure to either the .xyz or .int files described above. For example, the vibrational analysis program can save the tenth normal mode in sample.010. A molecular dynamics-based program might save its tenth 0.1 picosecond frame (or an energy minimizer its tenth partially minimized intermediate) in a file of the same name.

SAMPLE.LOG

The Force Field Explorer interface to TINKER saves results of all calculations launched from the GUI to a log file with the .log suffix. Any output that would normally be directed to the screen after starting a program from the command line is appended to this log file by Force Field Explorer.

SAMPLE.ARC

A TINKER archive file is simply a series of .xyz Cartesian coordinate files appended together one after another. This file can be used to condense the results from intermediate stages of an optimization, frames from a molecular dynamics trajectory, or set of normal mode vibrations into a single file for storage. TINKER archive files can be displayed as ``movies'' by the Force Field Explorer modeling program.

SAMPLE.PDB

This file type contains coordinate information in the PDB format developed by the Brookhaven Protein Data Bank for deposition of model structures based on macromolecular X-ray diffraction and NMR data. Although TINKER itself does not use .pdb files directly for input/output, auxiliary programs are provided with the system for interconverting .pdb files with the .xyz format described above.

SAMPLE.SEQ

This file type contains the primary sequence of a biopolymer in the standard one-letter code with 50 residues per line. The .seq file for a biopolymer is generated automatically when a PDB file is converted to TINKER .xyz format or when using the PROTEIN or NUCLEIC programs to build a structure from sequence It is required for the reverse conversion of a TINKER file back to PDB format..

SAMPLE.FRAC

The fractional coordinates corresponding to the asymmetric unit of a crystal unit cell are stored in the .frac file. The internal format of this file is identical to the .xyz file; except that the coordinates are fractional instead of in Angstrom units.

SAMPLE.XMOL

The ARCHIVE program has the option of converting a series of .xyz cycle files into an XMakemol XYZ file. These files can be displayed as a movie using the XMakemol display program. Note that the .xmol file format does not contain TINKER atom type information, so it is not possible to convert an .xmol file back into a TINKER .xyz file.

SAMPLE.CAR

The ARCHIVE program has the option of converting a series of .xyz cycle files into an Accelerys InsightII coordinate archive file. These files can be displayed as a movie using the InsightII display program. Note that the .car file format does not contain TINKER atom type information, so it is not possible to convert a .car file back into a TINKER .XYZ file.

PARAMETER FILES

The potential energy parameter files distributed with the TINKER package all end in the extension .prm, although this is not required by the programs themselves. Each of these files contains a definition of the potential energy functional forms for that force field as well as values for individual energy parameters. For example, the mm3pro.prm file contains the energy parameters and definitions needed for a protein-specific version of the MM3 force field.

This section of the manual contains a brief description of each of the TINKER potential energy programs. A detailed example showing how to run each program is included in a later section. The programs listed below are all part of the main, supported distribution. Additional source code for various unsupported programs can be found in the /other directory of the TINKER distribution.

ALCHEMY

A simple program to perform very basic free energy perturbation calculations. This program is provided mostly for demonstration purposes. For example, we use ALCHEMY in a molecular modeling course laboratory exercise to perform such classic mutations as chloride to bromide and ethane to methanol in water. The present version uses the perturbation formula and windowing with an explicit mapping of atoms involved in the mutation (``AMBER''-style), instead of thermodynamic integration and independent freely propagating groups of mutated atoms (``CHARMM''-style). Some of the code specific to this program is limited to the AMBER and OPLS potential functional forms, but could be easily generalized to handle other potentials. A more general and sophisticated version is currently under development.

ANALYZE

Provides information about a specific molecular structure. The program will ask for the name of a structure file, which must be in the TINKER .xyz file format, and the type of analysis desired. Options allow output of: (1) total potential energy of the system, (2) breakdown of the energy by potential function type or over individual atoms, (3) computation of the total dipole moment and its components, moments of inertia and radius of gyration, (4) listing of the parameters used to compute selected interaction energies, (5) energies associated with specified individual interactions.

ANNEAL

Performs a molecular dynamics simulated annealing computation. The program starts from a specified input molecular structure in TINKER .xyz format. The trajectory is updated using either a modified Beeman or a velocity Verlet integration method. The annealing protocol is implemented by allowing smooth changes between starting and final values of the system temperature via the Groningen method of coupling to an external bath. The scaling can be linear or sigmoidal in nature. In addition, parameters such as cutoff distance can be transformed along with the temperature. The user must input the desired number of dynamics steps for both the equilibration and cooling phases, a time interval for the dynamics steps, and an interval between coordinate/trajectory saves. All saved coordinate sets along the trajectory are placed in sequentially numbered cycle files.

DYNAMIC

Performs a molecular dynamics (MD) or stochastic dynamics (SD) computation. Starts either from a specified input molecular structure (an .xyz file) or from a structure-velocity-acceleration set saved from a previous dynamics trajectory (a restart from a .dyn file). MD trajectories are propagated using either a modified Beeman or a velocity Verlet integration method. SD is implemented via our own derivation of a velocity Verlet-based algorithm. In addition the program can perform full crystal calculations, and can operate in constant energy mode or with maintenance of a desired temperature and/or pressure using the Groningen method of coupling to external baths. The user must input the desired number of dynamics steps, a time interval for the dynamics steps, and an interval between coordinate/trajectory saves. Coordinate sets along the trajectory can be saved as sequentially numbered cycle files or directly to a TINKER archive .arc file. At the same time that a point along the trajectory is saved, the complete information needed to restart the trajectory from that point is updated and stored in the .dyn file.

GDA

A program to implement Straub's Gaussian Density Annealing algorithm over an effective series of analytically smoothed potential energy surfaces. This method can be viewed as an extended stochastic version of the diffusion equation method of Scheraga, et al., and also has many similar features to the TINKER Potential Smoothing and Search (PSS) series of programs. The current version of GDA is similar to but does not exactly reproduce Straub's published method and is limited to argon clusters and other simple systems involving only van der Waals interactions; further modification and development of this code is currently underway in the Ponder research group. As with other programs involving potential smoothing, GDA currently requires use of the smooth.prm force field parameters.

MINIMIZE

The MINIMIZE program performs a limited memory L-BFGS minimization of an input structure over Cartesian coordinates using a modified version of the algorithm of Jorge Nocedal. The method requires only the potential energy and gradient at each step along the minimization pathway. It requires storage space proportional to the number of atoms in the structure. The MINIMIZE procedure is recommended for preliminary minimization of trial structures to an rms gradient of 1.0 to 0.1 kcal/mole/Å. It has a relatively fast cycle time and is tolerant of poor initial structures, but converges in a slow, linear fashion near the minimum. The user supplies the name of the TINKER .xyz coordinates file and a target rms gradient value at which the minimization will terminate. Output consists of minimization statistics written to the screen or redirected to an output file, and the new coordinates written to updated .xyz files or to cycle files.

MINIROT

The MINIROT program uses the same limited memory L-BFGS method as MINIMIZE, but performs the computation in terms of dihedral angles instead of Cartesian coordinates. Output is saved in an updated .int file or in cycle files.

MINRIGID

The MINRIGID program is similar to MINIMIZE except that it operates on rigid bodies starting from a TINKER .xyz coordinate file and the rigid body group definitions found in the corresponding .key file. Output is saved in an updated .xyz file or in cycle files.

MONTE

The MONTE program implements the Monte Carlo Minimization algorithm developed by Harold Scheraga's group and others. The procedure takes Monte Carlo steps for either a single atom or a single torsional angle, then performs a minimization before application of the Metropolis sampling method. This results in effective sampling of a modified potential surface where the only possible energy levels are those of local minima on the original surface. The program can be easily modified to elaborate on the available move set.

NEWTON

A truncated Newton minimization method which requires potential energy, gradient and Hessian information. This procedure has significant advantages over standard Newton methods, and is able to minimize very large structures completely. Several options are provided with respect to minimization method and preconditioning of the Newton equations. The default options are recommended unless the user is familiar with the math involved. This program operates in Cartesian coordinate space and is fairly tolerant of poor input structures. Typical algorithm iteration times are longer than with nonlinear conjugate gradient or variable metric methods, but many fewer iterations are required for complete minimization. NEWTON is usually the best choice for minimizations to the 0.01 to 0.000001 kcal/mole/Å level of rms gradient convergence. Tests for directions of negative curvature can be removed, allowing NEWTON to be used for optimization to conformational transition state structures (this only works if the starting point is very close to the transition state). Input consists of a TINKER .xyz coordinates file; output is an updated set of minimized coordinates and minimization statistics.

NEWTROT

The NEWTROT program is similar to NEWTON except that it requires a .int file as input and then operates in terms of dihedral angles as the minimization variables. Since the dihedral space Hessian matrix of an arbitrary structure is often indefinite, this method will often not perform as well as the other, simpler dihedral angle based minimizers.

OPTIMIZE

The OPTIMIZE program performs a optimally conditioned variable metric minimization of an input structure over Cartesian coordinates using an algorithm due to William Davidon. The method does not perform line searches, but requires computation of energies and gradients as well as storage for an estimate of the inverse Hessian matrix. The program operates on Cartesian coordinates from a TINKER .xyz file. OPTIMIZE will typically converge somewhat faster and more completely than MINIMIZE. However, the need to store and manipulate a full inverse Hessian estimate limits its use to structures containing less than a few hundred atoms on workstation class machines. As with the other minimizers, OPTIMIZE needs input coordinates and an rms gradient cutoff criterion. The output coordinates are saved in updated .xyz files or as cycle files.

OPTIROT

The OPTIROT program is similar to OPTIMIZE except that it operates on dihedral angles starting from a TINKER .int internal coordinate file. This program is usually the preferred method for most dihedral angle optimization problems since Truncated Newton methods appear, in our hands, to lose some of their efficacy in moving from Cartesian to torsional coordinates.

OPTRIGID

The OPTRIGID program is similar to OPTIMIZE except that it operates on rigid bodies starting from a TINKER .xyz coordinate file and the rigid body atom group definitions found in the corresponding .key file. Output is saved in an updated .xyz file or in cycle files.

PATH

A program that implements a variant of Elber's Lagrangian multiplier-based reaction path following algorithm. The program takes as input a pair of structural minima as TINKER .xyz files, and then generates a user specified number of points along a path through conformational space connecting the input structures. The intermediate structures are output as TINKER cycle files, and the higher energy intermediates can be used as input to a Newton-based optimization to locate conformational transition states.

PSS

Implements our version of a potential smoothing and search algorithm for the global optimization of molecular conformation. An initial structure in .xyz format is first minimized in Cartesian coordinates on a series of increasingly smoothed potential energy surfaces. Then the smoothing procedure is reversed with minimization on each successive surface starting from the coordinates of the minimum on the previous surface. A local search procedure is used during the backtracking to explore for alternative minima better than the one found during the current minimization. The final result is usually a very low energy conformation or, in favorable cases, the global energy minimum conformation. The minimum energy coordinate sets found on each surface during both the forward smoothing and backtracking procedures are placed in sequentially numbered cycle files.

PSSRIGID

This program implements the potential smoothing and search method as described above for the PSS program, but performs the computation in terms of keyfile-defined rigid body atom groups instead of Cartesian coordinates. Output is saved in numbered cycle files with the .xyz file format.

PSSROT

This program implements the potential smoothing and search method as described above for the PSS program, but performs the computation in terms of a set of user-specified dihedral angles instead of Cartesian coordinates. Output is saved in numbered cycle files with the .int file format.

SADDLE

A program for the location of a conformational transition state between two potential energy minima. SADDLE uses a conglomeration of ideas from the Bell-Crighton quadratic path and the Halgren-Lipscomb synchronous transit methods. The basic idea is to perform a nonlinear conjugate gradient optimization in a subspace orthogonal to a suitably defined reaction coordinate. The program requires as input the coordinates (TINKER .xyz files) of the two minima and an rms gradient convergence criterion for the optimization. The current estimate of the transition state structure is written to the file TSTATE.XYZ. Crude transition state structures generated by SADDLE can sometimes be refined using the NEWTON program. Optionally, a scan of the interconversion pathway can be made at each major iteration.

SCAN

A program for general conformational search of an entire potential energy surface via a basin hopping method. The program takes as input a TINKER .xyz coordinates file which is then minimized to find the first local minimum for a search list. A series of activations along various normal modes from this initial minimum are used as seed points for additional minimizations. Whenever a previously unknown local minimum is located it is added to the search list. When all minima on the search list have been subjected to the normal mode activation without locating additional new minima, the program terminates. The individual local minima are written to cycle files as they are discovered. While the SCAN program can be used on standard undeformed potential energy surfaces, we have found it to be most useful for quickly ``scanning'' a smoothed energy surface to enumerate the major basins of attraction spaning the entire surface.

SNIFFER

A program that implements the Sniffer global optimization algorithm of Butler and Slaminka, a discrete version of Griewank's global search trajectory method. The program takes an input TINKER .xyz coordinates file and shakes it vigorously via a modified dynamics trajectory before, hopefully, settling into a low lying minimum. Some trial and error is often required as the current implementation is sensitive to various parameters and tolerances that govern the computation. At present, these parameters are not user accessible, and must be altered in the source code. However, this method can do a good job of quickly optimizing conformation within a limited range of convergence.

TESTGRAD

The TESTGRAD program computes and compares the analytical and numerical first derivatives (i.e., the gradient vector) of the potential energy for a Cartesian coordinate input structure. The output can be used to test or debug the current potential or any added user defined energy terms.

TESTHESS

The TESTHESS program computes and compares the analytical and numerical second derivatives (i.e., the Hessian matrix) of the potential energy for a Cartesian coordinate input structure. The output can be used to test or debug the current potential or any added user defined energy terms.

TESTLIGHT

A program to compare the efficiency of different nonbonded neighbor methods for the current molecular system. The program times the computation of energy and gradient for the van der Waals and charge-charge electrostatic potential terms using a simple double loop over all interactions and using the Method of Lights algorithm to select neighbors. The results can be used to decide whether the Method of Lights has any CPU time advantage for the current structure. Both methods should give exactly the same answer in all cases, since the identical individual interactions are computed by both methods. The default double loop method is faster when cutoffs are not used, or when the cutoff sphere contains about half or more of the total system of unit cell. In cases where the cutoff sphere is much smaller than the system size, the Method of Lights can be much faster since it avoids unnecessary calculation of distances beyond the cutoff range.

TESTROT

The TESTROT program computes and compares the analytical and numerical first derivatives (i.e., the gradient vector) of the potential energy with respect to dihedral angles. Input is a TINKER .int internal coordinate file. The output can be used to test or debug the current potential functions or any added user defined energy terms.

TIMER

A simple program to provide timing statistics for energy function calls within the TINKER package. TIMER requires an input .xyz file and outputs the CPU time (wall clock time on some machine types) needed to perform a specified number of energy, gradient and Hessian evaluations.

TIMEROT

This program is similar to TIMER, only it operates over dihedral angles via input of a TINKER .int internal coordinate file. In the current version, the torsional Hessian is computed numerically from the analytical torsional gradient.

VIBRATE

A program to perform vibrational analysis by computing and diagonalizing the full Hessian matrix (i.e., the second partial derivatives) for an input structure (a TINKER .xyz file). Eigenvalues and eigenvectors of the mass weighted Hessian (i.e., the vibrational frequencies and normal modes) are also calculated. Structures corresponding to individual normal mode motions can be saved in cycle files.

VIBROT

The program VIBROT forms the torsional Hessian matrix via numerical differentiation of the analytical torsional gradient. The Hessian is then diagonalized and the eigenvalues are output. The present version does not compute the kinetic energy matrix elements needed to convert the Hessian into the torsional normal modes; this will be added in a later version. The required input is a TINKER .int internal coordinate file.

XTALFIT

The XTALFIT program is of use in the automated fitting of potential parameters to crystal structure and thermodynamic data. XTALFIT takes as input several crystal structures (TINKER .xyz files with unit cell parameters in corresponding keyfiles) as well as information on lattice energies and dipole moments of monomers. The current version uses a nonlinear least squares optimization to fit van der Waals and electrostatic parameters to the input data. Bounds can be placed on the values of the optimization parameters.

XTALMIN

A program to perform full crystal minimizations. The program takes as input the structure coordinates and unit cell lattice parameters. It then alternates cycles of Newton-style optimization of the structure and conjugate gradient optimization of the crystal lattice parameters. This alternating minimization is slower than more direct optimization of all parameters at once, but is somewhat more robust in our hands. The symmetry of the original crystal is not enforced, so interconversion of crystal forms may be observed in some cases.

This section of the manual contains a brief description of each of the TINKER structure manipulation, geometric calculation and auxiliary programs. A detailed example showing how to run each program is included in a later section. The programs listed below are all part of the main, supported distribution. Additional source code for various unsupported programs can be found in the /other directory of the TINKER distribution.

ARCHIVE

A program for concatenating TINKER cycle files into a single archive file; useful for storing the intermediate results of minimizations, dynamics trajectories, and so on. The archive file can be written in TINKER format, or in formats usable with MSI's InsightII (their CAR file with .msi extension) or with XMakemol (their file format with .xmol extension). Only active atoms are written into the InsightII and XMakemol output files, allowing display of partial structures. The program can also extract individual cycle files from a TINKER archive.

CORRELATE

A program to compute time correlation functions from collections of TINKER cycle files. Its use requires a user supplied function property that computes the value of the property for which a time correlation is desired for two input structures. A sample routine is supplied that computes either a velocity autocorrelation function or an rms structural superposition as a function of time. The main body of the program organizes the overall computation in an efficient manner and outputs the final time correlation function.

CRYSTAL

A program for the manipulation of crystal structures including interconversion of fractional and Cartesian coordinates, generation of the unit cell from an asymmetric unit, and building of a crystalline block of specified size via replication of a single unit cell. The present version can handle about 25 of the most common space groups, others can easily be added as needed by modification of the routine symmetry.

DIFFUSE

A program to compute the self-diffusion constant for a homogeneous liquid via the Einstein equation. A previously saved dynamics trajectory is read in and ``unfolded'' to reverse translation of molecules due to use of periodic boundary conditions. The average motion over all molecules is then used to compute the self-diffusion constant. While the current program assumes a homogeneous system, it should be easy to modify the code to handle diffusion of individual molecules or other desired effects.

DISTGEOM

A program to perform distance geometry calculations using variations on the classic metric matrix method. A user specified number of structures consistent with keyfile input distance and dihedral restraints is generated. Bond length and angle restraints are derived from the input structure. Trial distances between the triangle smoothed lower and upper bounds can be chosen via any of several metrization methods, including a very effective partial random pairwise scheme. The correct radius of gyration of the structure is automatically maintained by choosing trial distances from Gaussian distributions of appropriate mean and width. The initial embedded structures can be further refined against a geometric restraint-only potential using either a sequential minimization protocol or simulated annealing.

DOCUMENT

The DOCUMENT program is provided as a minimal listing and documentation tool. It operates on the TINKER source code, either individual files or the complete source listing produced by the command script listing.make, to generate lists of routines, common blocks or valid keywords. In addition, the program has the ability to output a formatted parameter listing from the standard TINKER parameter files.

INTEDIT

A program to allow interactive inspection and alteration of the internal coordinate definitions and values of a TINKER structure. If the structure is altered, the user has the option to write out a new internal coordinates file upon exit.

INTXYZ

A program to convert a TINKER .int internal coordinates formatted file into a TINKER .xyz Cartesian coordinates formatted file.

NUCLEIC

A program for automated building of nucleic acid structures. Upon interactive input of a nucleotide sequence with optional phosphate backbone angles, the program builds internal and Cartesian coordinates. Standard bond lengths and angles are used. Both DNA and RNA sequences are supported as are A-, B- and Z-form structures. Double helixes of complementary sequence can be automatically constructed via a rigid docking of individual strands.

PDBXYZ

A program for converting a Brookhaven Protein Data Bank file (a PDB file) into a TINKER .xyz Cartesian coordinate file. If the PDB file contains only protein/peptide amino acid residues, then standard protein connectivity is assumed, and transferred to the .xyz file. For non-protein portions of the PDB file, atom connectivity is determined by the program based on interatomic distances. The program also has the ability to add or remove hydrogen atoms from a protein as required by the force field specified during the computation.

POLARIZE

A program for computing molecular polarizability from an atom-based distributed model of polarizability. A damped interaction model due to Thole is optionally via keyfile settings. A TINKER .xyz file is required as input. The output consists of the overall polarizability tensor in the global coordinates and its eigenvalues.

PRMEDIT

A program for formatting and renumbering TINKER force field parameter files. When atom types or classes are added to a parameter file, this utility program has the ability to renumber all the atom records sequentially, and alter type and class numbers in all other parameter entries to maintain consistency.

PROTEIN

A program for automated building of peptide and protein structures. Upon interactive input of an amino acid sequence with optional phi/psi/omega/chi angles, D/L chirality, etc., the program builds internal and Cartesian coordinates. Standard bond lengths and angles are assumed for the peptide. The program will optionally convert the structure to a cyclic peptide, or add either or both N- and C-terminal capping groups. Atom type numbers are automatically assigned for the specified force field. The final coordinates and a sequence file are produced as the output.

RADIAL

A program to compute the pair radial distribution function between two atom types. The user supplies the two atom names for which the distribution function is to be computed, and the width of the distance bins for data analysis. A previously saved dynamics trajectory is read as input. The raw radial distribution and a spline smoothed version are then output from zero to a distance equal to half the minimum periodic box dimension. The atom names are matched to the atom name column of the TINKER .xyz file, independent of atom type.

SPACEFILL

A program to compute the volume and surface areas of molecules. Using a modified version of Connolly's original analytical description of the molecular surface, the program determines either the van der Waals, accessible or molecular (contact/reentrant) volume and surface area. Both surface area and volume are broken down into their geometric components, and surface area is decomposed into the convex contribution for each individual atom. The probe radius is input as a user option, and atomic radii can be set via the keyword file. If TINKER archive files are used as input, the program will compute the volume and surface area of each structure in the input file.

SPECTRUM

A program to compute a power spectrum from velocity autocorrelation data. As input, this program requires a velocity autocorrelation function as produced by the CORRELATE program. This data, along with a user input time step, are Fourier transformed to generate the spectral intensities over a wavelength range. The result is a power spectrum, and the positions of the bands are those predicted for an infrared or Raman spectrum. However, the data is not weighted by molecular dipole moment derivatives as would be required to produce correct IR intensities.

SUPERPOSE

A program to superimpose two molecular structures in 3-dimensions. A variety of options for input of the atom sets to be used during the superposition are presented interactively to the user. The superposition can be mass-weighted if desired, and the coordinates of the second structure superimposed on the first structure are optionally output. If TINKER archive files are used as input, the program will compute all pairwise superpositions between structures in the input files.

SYBYLXYZ

A program for converting a TRIPOS Sybyl MOL2 file into a TINKER .xyz Cartesian coordinate file. The current version of the program does not attempt to convert the Sybyl atoms types into the active TINKER force field types, i.e., all atoms types are simply set to zero.

TVIEW

This is a molecule viewing program derived from the well-know Rasmol program of Roger Sayle. TVIEW is modified to remove most of the protein-specific options and to directly read the TINKER .xyz file format. The original RasMol program has been altered to allow selection and specification by atoms instead of residues. We hope to provide additional functionality in future versions of TVIEW, especially the ability to animate the viewing of sequences of coordinate snapshots from a minimization or dynamic trajectory.

XYZEDIT

A program that performs and of a variety of manipulations on an input TINKER .xyz Cartesian coordinates formatted file. The present version of the program has the following interactively selectable options: (1) Offset the Numbers of the Current Atoms, (2) Deletion of Individual Specified Atoms, (3) Deletion of Specified Types of Atoms, (4) Deletion of Atoms outside Cutoff Range, (5) Insertion of Individual Specified Atoms, (6) Replace Old Atom Type with a New Type, (7) Assign Connectivities based on Distance, (8) Convert Units from Bohrs to Angstroms, (9) Invert thru Origin to give Mirror Image, (10) Translate Center of Mass to the Origin, (11) Translate a Specified Atom to the Origin, (12) Translate and Rotate to Inertial Frame, (13) Move to Specified Rigid Body Coordinates, (14) Create and Fill a Periodic Boundary Box, (15) Soak Current Molecule in Box of Solvent, (16) Append another XYZ file to Current One. In most cases, multiply options can be applied sequentially to an input file. At the end of the editing process, a new version of the original .xyz file is written as output.

XYZINT

A program for converting a TINKER .xyz Cartesian coordinate formatted file into a TINKER .int internal coordinates formatted file. This program can optionally use an existing internal coordinates file as a template for the connectivity information.

XYZPDB

A program for converting a TINKER .xyz Cartesian coordinate file into a Brookhaven Protein Data Bank file (a PDB file).

XYZSYBYL

A program to convert a TINKER .xyz Cartesian coordinates file into a TRIPOS Sybyl MOL2 file. The conversion generates only the MOLECULE, ATOM, BOND and SUBSTRUCTURE record type in the MOL2 file. Generic Sybyl atom types are used in most cases; while these atom types may need to be altered in some cases, Sybyl is usually able to correctly display the resulting MOL2 file.

The TINKER package is distributed with several force field parameter sets, implementing a selection of widely used literature force fields as well as the TINKER force field currently under construction in the Ponder lab. We try to exactly reproduce the intent of the original authors of our distributed, third-party force fields. In all cases the parameter sets have been validated against literature reports, results provided by the original developers, or calculations made with the authentic programs. With the few exceptions noted below, TINKER calculations can be treated as authentic results from the genuine force fields. A brief description of each parameter set, including some still in preparation and not distributed with the current version, is provided below with lead literature references for the force field:

AMOEBA.PRM

Parameters for the AMOEBA polarizable atomic multipole force field. As of the current TINKER release, we have completed parametrization for a number of ions and small organic molecules. For further information, or if you are interested in developing or testing parameters for other small molecules, please contact the Ponder lab.

P. Ren and J. W. Ponder, A Consistent Treatment of Inter- and Intramolecular Polarization in Molecular Mechanics Calculations, J. Comput. Chem., 23, 1497-1506 (2002)

P. Ren and J. W. Ponder, Polarizable Atomic Multipole Water Model for Molecular Mechanics Simulation, J. Phys. Chem. B, 107, 5933-5947 (2003)

P. Ren and J. W. Ponder, Ion Solvation Thermodynamics from Simulation with a Polarizable Force Field, A. Grossfield, J. Am. Chem. Soc., 125, 15671-15682 (2003)

AMOEBAPRO.PRM

Preliminary protein parameters for the AMOEBA polarizable atomic multipole force field. While the distributed parameters are still subject to minor alteration as we continue validation, they are now stable enough for other groups to begin using them. For further information, or if you are interested in testing the protein parameter set, please contact the Ponder lab.

J. W. Ponder and D. A. Case, Force Fields for Protein Simulation, Adv. Prot. Chem., 66, 27-85 (2003)

P. Ren and J. W. Ponder, Polarizable Atomic Multipole-based Potential for Proteins: Model and Parameterization, in preparation

AMBER94.PRM

AMBER ff94 parameters for proteins and nucleic acids. Note that with their ``Cornell'' force field, the Kollman group has devised separate, fully independent partial charge values for each of the N- and C-terminal amino acid residues. At present, the terminal residue charges for TINKER's version maintain the correct formal charge, but redistributed somewhat at the alpha carbon atoms from the original Kollman group values. The total magnitude of the redistribution is less than 0.01 electrons in most cases.

W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz, Jr., D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell and P. A. Kollman, A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids, and Organic Molecules, J. Am. Chem. Soc., 117, 5179-5197 (1995) [ff94]

G. Moyna, H. J. Williams, R. J. Nachman and A. I. Scott, Conformation in Solution and Dynamics of a Structurally Constrained Linear Insect Kinin Pentapeptide Analogue, Biopolymers, 49, 403-413 (1999) [AIB charges]

W. S. Ross and C. C. Hardin, Ion-Induced Stabilization of the G-DNA Quadruplex: Free Energy Perturbation Studies, J. Am. Chem. Soc., 116, 4363-4366 (1994) [alkali metal ions]

J. Aqvist, Ion-Water Interaction Potentials Derived from Free Energy Perturbation Simulations, J. Phys. Chem., 94, 8021-8024, 1990 [alkaline earth Ions, radii adapted for Amber combining rule]

Current force field parameter values and suggested procedures for development of parameters for additional molecules are available from the Amber web site in the Case lab at Scripps, http://amber.scripps.edu/

AMBER96.PRM

AMBER ff96 parameters for proteins and nucleic acids. The only change from the ff94 parameter set is in the torsional parameters for the protein phi/psi angles. These values were altered to give better agreement with changes of ff96 with LMP2 QM results from the Friesner lab on alanine dipeptide and tetrapeptide.

P. Kollman, R. Dixon, W. Cornell, T. Fox, C. Chipot and A. Pohorille, The Development/ Application of a 'Minimalist' Organic/Biochemical Molecular Mechanic Force Field using a Combination of ab Initio Calculations and Experimental Data, in Computer Simulation of Biomolecular Systems, W. F. van Gunsteren, P. K. Weiner, A. J. Wilkinson, eds., Volume 3, 83-96 (1997) [ff96]

Current force field parameter values and suggested procedures for development of parameters for additional molecules are available from the Amber web site in the Case lab at Scripps, http://amber.scripps.edu/

AMBER98.PRM

AMBER ff98 parameters for proteins and nucleic acids. The only change from the ff94 parameter set is in the glycosidic torsional parameters that control sugar pucker.

T. E. Cheatham III, P. Cieplak and P. A. Kollman, A Modified Version of the Cornell et al. Force Field with Improved Sugar Pucker Phases and Helical Repeat, J. Biomol. Struct. Dyn., 16, 845-862 (1999)

Current force field parameter values and suggested procedures for development of parameters for additional molecules are available from the Amber web site in the Case lab at Scripps, http://amber.scripps.edu/

AMBER99.PRM

AMBER ff99 parameters for proteins and nucleic acids. The original partial charges from the ff94 parameter set are retained, but many of the bond, angle and torsional parameters have been revised to provide better general agreement with experiment.

J. Wang, P. Cieplak and P. A. Kollman, How Well Does a Restrained Electrostatic Potential (RESP) Model Perform in Calcluating Conformational Energies of Organic and Biological Molecules?, J. Comput. Chem., 21, 1049-1074 (2000)

Current force field parameter values and suggested procedures for development of parameters for additional molecules are available from the Amber web site in the Case lab at Scripps, http://amber.scripps.edu/

CHARMM19.PRM

CHARMM19 united-atom parameters for proteins. The nucleic acid parameter are not yet implemented. There are some differences between authentic CHARMM19 and the TINKER version due to replacement of CHARMM impropers by torsions for cases that involve atoms not bonded to the trigonal atom and TINKER's use of all possible torsions across a bond instead of a single torsion per bond.

E. Neria, S. Fischer and M. Karplus, Simulation of Activation Free Energies in Molecular Systems, J. Chem. Phys., 105, 1902-1921 (1996)

L. Nilsson and M. Karplus, Empirical Energy Functions for Energy Minimizations and Dynamics of Nucleic Acids, J. Comput. Chem., 7, 591-616 (1986)

W. E. Reiher III, Theoretical Studies of Hydrogen Bonding, Ph.D. Thesis, Department of Chemistry, Harvard University, Cambridge, MA, 1985

CHARMM27.PRM

CHARMM27 all-atom parameters for proteins and lipids. Most of the nucleic acid and small model compound parameters are not yet implemented. We plan to provide these additional parameters in due course.

N. Foloppe and A. D. MacKerell, Jr., All-Atom Empirical Force Field for Nucleic Acids: 1) Parameter Optimization Based on Small Molecule and Condensed Phase Macromolecular Target Data, J. Comput. Chem., 21, 86-104 (2000) [CHARMM27]

N. Banavali and A. D. MacKerell, Jr., All-Atom Empirical Force Field for Nucleic Acids: 2) Application to Molecular Dynamics Simulations of DNA and RNA in Solution, J. Comput. Chem., 21, 105-120 (2000)

A. D. MacKerrell, Jr., et al., All-Atom Empirical Potential for Molecular Modeling and Dynamics Studies of Proteins, J. Phys. Chem. B, 102, 3586-3616 (1998) [CHARMM22]

A. D. MacKerell, Jr., J. Wiorkeiwicz-Kuczera and M. Karplus, An All-Atom Empirical Energy Function for the Simulation of Nucleic Acids, J. Am. Chem. Soc., 117, 11946-11975 (1995)

S. E. Feller, D. Yin, R. W. Pastor and A. D. MacKerell, Jr., Molecular Dynamics Simulation of Unsaturated Lipids at Low Hydration: Parametrization and Comparison with Diffraction Studies, Biophysical Journal, 73, 2269-2279 (1997) [alkenes]

R. H. Stote and M. Karplus, Zinc Binding in Proteins and Solution - A Simple but Accurate Nonbonded Representation, Proteins, 23, 12-31 (1995) [zinc ion]

Current and legacy parameter values are available from the CHARMM force field web site on Alex MacKerell's Research Interests page at the University of Maryland School of Pharmacy, https://rxsecure.umaryland.edu/research/amackere/research.html/

DUDEK.PRM

Protein-only parameters for the early 1990's TINKER force field with multipole values of Dudek and Ponder. The current file contains only the multipole values from the 1995 paper by Dudek and Ponder. This set is now superceeded by the more recent TINKER force field developed by Pengyu Ren (see WATER.PRM, below).

M. J. Dudek and J. W. Ponder, Accurate Electrostatic Modelling of the Intramolecular Energy of Proteins, J. Comput. Chem., 16, 791-816 (1995)

ENCAD.PRM

ENCAD parameters for proteins and nucleic acids. (in preparation)

M. Levitt, M. Hirshberg, R. Sharon and V. Daggett, Potential Energy Function and Parameters for Simulations of the Molecular Dynamics of Protein and Nucleic Acids in Solution, Comp. Phys. Commun., 91, 215-231 (1995)

M. Levitt, M. Hirshberg, R. Sharon, K. E. Laidig and V. Daggett, Calibration and Testing of a Water Model for Simulation of the Molecular Dynamics of Protein and Nucleic Acids in Solution, J. Phys. Chem. B, 101, 5051-5061 (1997) [F3C water]

HOCH.PRM

Simple NMR-NOE force field of Hoch and Stern.

J. C. Hoch and A. S. Stern, A Method for Determining Overall Protein Fold from NMR Distance Restraints, J. Biomol. NMR, 2, 535-543 (1992)

MM2.PRM

Full MM2(1991) parameters including p-systems. The anomeric and electronegativity correction terms included in some later versions of MM2 are not implemented.

N. L. Allinger, Conformational Analysis. 130. MM2. A Hydrocarbon Force Field Utilizing V1 and V2 Torsional Terms, J. Am. Chem. Soc., 99, 8127-8134 (1977)

J. T. Sprague, J. C. Tai, Y. Yuh and N. L. Allinger, The MMP2 Calculational Method, J. Comput. Chem., 8, 581-603 (1987)

J. C. Tai and N. L. Allinger, Molecular Mechanics Calculations on Conjugated Nitrogen-Containing Heterocycles, J. Am. Chem. Soc., 110, 2050-2055 (1988)

J. C. Tai, J.-H. Lii and N. L. Allinger, A Molecular Mechanics (MM2) Study of Furan, Thiophene, and Related Compounds, J. Comput. Chem., 10, 635-647 (1989)

N. L. Allinger, R. A. Kok and M. R. Imam, Hydrogen Bonding in MM2, J. Comput. Chem., 9, 591-595 (1988)

L. Norskov-Lauritsen and N. L. Allinger, A Molecular Mechanics Treatment of the Anomeric Effect, J. Comput. Chem., 5, 326-335 (1984)

All parameters distributed with TINKER are from the ``MM2 (1991) Parameter Set'', as provided by N. L. Allinger, University of Georgia

MM3.PRM

Full MM3(2000) parameters including pi-systems. The directional hydrogen bonding term and electronegativity bond length corrections are implemented, but the anomeric and Bohlmann correction terms are not implemented.

N. L. Allinger, Y. H. Yuh and J.-H. Lii, Molecular Mechanics. The MM3 Force Field for Hydrocarbons. 1, J. Am. Chem. Soc., 111, 8551-8566 (1989)

J.-H. Lii and N. L. Allinger, Molecular Mechanics. The MM3 Force Field for Hydrocarbons. 2. Vibrational Frequencies and Thermodynamics, J. Am. Chem. Soc., 111, 8566-8575 (1989)

J.-H. Lii and N. L. Allinger, Molecular Mechanics. The MM3 Force Field for Hydrocarbons. 3. The van der Waals' Potentials and Crystal Data for Aliphatic and Aromatic Hydrocarbons, J. Am. Chem. Soc., 111, 8576-8582 (1989)

N. L. Allinger, H. J. Geise, W. Pyckhout, L. A. Paquette and J. C. Gallucci, Structures of Norbornane and Dodecahedrane by Molecular Mechanics Calculations (MM3), X-ray Crystallography, and Electron Diffraction, J. Am. Chem. Soc., 111, 1106-1114 (1989) [stretch-torsion cross term]

N. L. Allinger, F. Li and L. Yan, Molecular Mechanics. The MM3 Force Field for Alkenes, J. Comput. Chem., 11, 848-867 (1990)

N. L. Allinger, F. Li, L. Yan and J. C. Tai, Molecular Mechanics (MM3) Calculations on Conjugated Hydrocarbons, J. Comput. Chem., 11, 868-895 (1990)

J.-H. Lii and N. L. Allinger, Directional Hydrogen Bonding in the MM3 Force Field. I, J. Phys. Org. Chem., 7, 591-609 (1994)

J.-H. Lii and N. L. Allinger, Directional Hydrogen Bonding in the MM3 Force Field. II, J. Comput. Chem., 19, 1001-1016 (1998)

All parameters distributed with TINKER are from the ``MM3 (2000) Parameter Set'', as provided by N. L. Allinger, University of Georgia, August 2000

MM3PRO.PRM

Protein-only version of the MM3 parameters.

J.-H. Lii and N. L. Allinger, The MM3 Force Field for Amides, Polypeptides and Proteins, J. Comput. Chem., 12, 186-199 (1991)

OPLSUA.PRM

Complete OPLS-UA with united-atom parameters for proteins and many classes of organic molecules. Explicit hydrogens on polar atoms and aromatic carbons.

W. L. Jorgensen and J. Tirado-Rives, The OPLS Potential Functions for Proteins. Energy Minimizations for Crystals of Cyclic Peptides and Crambin, J. Am. Chem. Soc., 110, 1657-1666 (1988) [peptide and proteins]

W. L. Jorgensen and D. L. Severance, Aromatic-Aromatic Interactions: Free Energy Profiles for the Benzene Dimer in Water, Chloroform, and Liquid Benzene, J. Am. Chem. Soc., 112, 4768-4774 (1990) [aromatic hydrogens]

S. J. Weiner, P. A. Kollman, D. A. Case, U. C. Singh, C. Ghio, G. Alagona, S. Profeta, Jr. and P. Weiner, A New Force Field for Molecular Mechanical Simulation of Nucleic Acids and Proteins, J. Am. Chem. Soc., 106, 765-784 (1984) [united-atom ``AMBER/OPLS'' local geometry]

S. J. Weiner, P. A. Kollman, D. T. Nguyen and D. A. Case, An All Atom Force Field for Simulations of Proteins and Nucleic Acids, J. Comput. Chem., 7, 230-252 (1986) [all-atom "AMBER/OPLS" local geometry]

L. X. Dang and B. M. Pettitt, Simple Intramolecular Model Potentials for Water, J. Phys. Chem., 91, 3349-3354 (1987) [flexible TIP3P and SPC water]

W. L. Jorgensen, J. D. Madura and C. J. Swenson, Optimized Intermolecular Potential Functions for Liquid Hydrocarbons, J. Am. Chem. Soc., 106, 6638-6646 (1984) [hydrocarbons]

W. L. Jorgensen, E. R. Laird, T. B. Nguyen and J. Tirado-Rives, Monte Carlo Simulations of Pure Liquid Substituted Benzenes with OPLS Potential Functions, J. Comput. Chem., 14, 206-215 (1993) [substituted benzenes]

E. M. Duffy, P. J. Kowalczyk and W. L. Jorgensen, Do Denaturants Interact with Aromatic Hydrocarbons in Water?, J. Am. Chem. Soc., 115, 9271-9275 (1993) [benzene, naphthalene, urea, guanidinium, tetramethyl ammonium]

W. L. Jorgensen and C. J. Swenson, Optimized Intermolecular Potential Functions for Amides and Peptides. Structure and Properties of Liquid Amides, J. Am. Chem. Soc., 106, 765-784 (1984) [amides]

W. L. Jorgensen, J. M. Briggs and M. L. Contreras, Relative Partition Coefficients for Organic Solutes form Fluid Simulations, J. Phys. Chem., 94, 1683-1686 (1990) [chloroform, pyridine, pyrazine, pyrimidine]

J. M. Briggs, T. B. Nguyen and W. L. Jorgensen, Monte Carlo Simulations of Liquid Acetic Acid and Methyl Acetate with the OPLS Potential Functions, J. Phys. Chem., 95, 3315-3322 (1991) [acetic acid, methyl acetate]

H. Liu, F. Muller-Plathe and W. F. van Gunsteren, A Force Field for Liquid Dimethyl Sulfoxide and Physical Properties of Liquid Dimethyl Sulfoxide Calculated Using Molecular Dynamics Simulation, J. Am. Chem. Soc., 117, 4363-4366 (1995) [dimethyl sulfoxide]

J. Gao, X. Xia and T. F. George, Importance of Bimolecular Interactions in Developing Empirical Potential Functions for Liquid Ammonia, J. Phys. Chem., 97, 9241-9246 (1993) [ammonia]

J. Aqvist, Ion-Water Interaction Potentials Derived from Free Energy Perturbation Simulations, J. Phys. Chem., 94, 8021-8024 (1990) [metal ions]

W. S. Ross and C. C. Hardin, Ion-Induced Stabilization of the G-DNA Quadruplex: Free Energy Perturbation Studies, J. Am. Chem. Soc., 116, 4363-4366 (1994) [alkali metal ions]

J. Chandrasekhar, D. C. Spellmeyer and W. L. Jorgensen, Energy Component Analysis for Dilute Aqueous Solutions of Li+, Na+, F-, and Cl- Ions, J. Am. Chem. Soc., 106, 903-910 (1984) [halide ions]

Most parameters distributed with TINKER are from ``OPLS and OPLS-AA Parameters for Organic Molecules, Ions, and Nucleic Acids'' as provided by W. L. Jorgensen, Yale University, October 1997

OPLSAA.PRM

OPLS-AA force field with all-atom parameters for proteins and many general classes of organic molecules.

W. L. Jorgensen, D. S. Maxwell and J. Tirado-Rives, Development and Testing of the OPLS All-Atom Force Field on Conformational Energetics and Properties of Organic Liquids, J. Am. Chem. Soc., 117, 11225-11236 (1996)

D. S. Maxwell, J. Tirado-Rives and W. L. Jorgensen, A Comprehensive Study of the Rotational Energy Profiles of Organic Systems by Ab Initio MO Theory, Forming a Basis for Peptide Torsional Parameters, J. Comput. Chem., 16, 984-1010 (1995)

W. L. Jorgensen and N. A. McDonald, Development of an All-Atom Force Field for Heterocycles. Properties of Liquid Pyridine and Diazenes, THEOCHEM-J. Mol. Struct., 424, 145-155 (1998)

N. A. McDonald and W. L. Jorgensen, Development of an All-Atom Force Field for Heterocycles. Properties of Liquid Pyrrole, Furan, Diazoles, and Oxazoles, J. Phys. Chem. B, 102, 8049-8059 (1998)

R. C. Rizzo and W. L. Jorgensen, OPLS All-Atom Model for Amines: Resolution of the Amine Hydration Problem, J. Am. Chem. Soc., 121, 4827-4836 (1999)

M. L. P. Price, D. Ostrovsky and W. L. Jorgensen, Gas-Phase and Liquid-State Properties of Esters, Nitriles, and Nitro Compounds with the OPLS-AA Force Field, J. Comput. Chem., 22, 1340-1352 (2001)

All parameters distributed with TINKER are from ``OPLS and OPLS-AA Parameters for Organic Molecules, Ions, and Nucleic Acids'' as provided by W. L. Jorgensen, Yale University, October 1997

OPLSAAL.PRM

An improved OPLS-AA parameter set for proteins in which the only change is a reworking of many of the backbone and sidechain torsional parameters to give better agreement with LMP2 QM calculations. This parameter set is also known as OPLS(2000).

G. A. Kaminsky, R. A. Friesner, J. Tirado-Rives and W. L. Jorgensen, Evaluation and Reparametrization of the OPLS-AA Force Field for Proteins via Comparison with Accurate Quantum Chemical Calculations on Peptides, J. Phys. Chem. B, 105, 6474-6487 (2001)

SMOOTH.PRM

Version of OPLS-UA for use with potential smoothing. Largely adapted largely from standard OPLS-UA parameters with modifications to the vdw and improper torsion terms.

R. V. Pappu, R. K. Hart and J. W. Ponder, Analysis and Application of Potential Energy Smoothing and Search Methods for Global Optimization, J. Phys, Chem. B, 102, 9725-9742 (1998) [smoothing modifications]

SMOOTHAA.PRM

Version of OPLS-AA for use with potential smoothing. Largely adapted largely from standard OPLS-AA parameters with modifications to the vdw and improper torsion terms.

R. V. Pappu, R. K. Hart and J. W. Ponder, Analysis and Application of Potential Energy Smoothing and Search Methods for Global Optimization, J. Phys, Chem. B, 102, 9725-9742 (1998) [smoothing modifications]

WATER.PRM

The AMOEBA water parameters for a polarizable atomic multipole electrostatics model. This model is equal or better to the best available water models for many bulk and cluster properties.

P. Ren and J. W. Ponder, A Polarizable Atomic Multipole Water Model for Molecular Mechanics Simulation, J. Phys. Chem. B, 107, 5933-5947 (2003)

P. Ren and J. W. Ponder, Ion Solvation Thermodynamics from Simulation with a Polarizable Force Field, A. Grossfield, J. Am. Chem. Soc., 125, 15671-15682 (2003)

P. Ren and J. W. Ponder, Temperature and Pressure Dependence of the AMOEBA Water Model, J. Phys. Chem. B, 108, xxxx-xxxx (2004)

An earlier version the AMOEBA water model is described in: Yong Kong, Multipole Electrostatic Methods for Protein Modeling with Reaction Field Treatment, Biochemistry & Molecular Biophysics, Washington University, St. Louis, August, 1997 [available from http://dasher.wustl.edu/ponder/]

This section contains a description of the keyword parameters which may be used to define or alter the course of a TINKER calculation. The keyword control file is optional in the sense that all of the TINKER programs will run in the absence of a keyfile and will simply use default values or query the user for needed information. However, the keywords allow use of a wide variety of algorithmic and procedural options, many of which are unavailable interactively.

Keywords are read from the keyword control file. All programs look first for a keyfile with the same base name as the input molecular system and ending in the extension .key. If this file does not exist, then TINKER tries to use a generic keyfile with the name tinker.key and located in the same directory as the input system. If neither a system specific nor a generic keyfile is present, TINKER will continue by using default values for keyword options and asking interactive questions as necessary.

TINKER searches the keyfile during the course of a calculation for relevant keywords that may be present. All keywords must appear as the first word on the line. Any blank space to the left of the keyword is ignored, and all contents of the keyfiles are case insensitive. Some keywords take modifiers; i.e., TINKER looks further on the same line for additional information, such as the value of some parameter related to the keyword. Modifier information is read in free format, but must be completely contained on the same line as the original keyword. Any lines contained in the keyfile which do not qualify as valid keyword lines are treated as comments and are simply ignored.

Several keywords take a list of integer values (atom numbers, for example) as modifiers. For these keywords the integers can simply be listed explicitly and separated by spaces, commas or tabs. If a range of numbers is desired, it can be specified by listing the negative of the first number of the range, followed by a separator and the last number of the range. For example, the keyword line ACTIVE 4 -9 17 23 could be used to add atoms 4, 9 through 17, and 23 to the set of active atoms during a TINKER calculation.

Listed below are the valid TINKER keywords sorted into groups by general function. The section ends with an alphabetical listing of the individual keywords along with brief descriptions of their action and possible modifiers, and examples of usage.

Keywords Grouped by Functionality

OUTPUT CONTROL KEYWORDS

ARCHIVE DEBUG DIGITS

ECHO EXIT-PAUSE NOVERSION OVERWRITE PRINTOUT SAVE-CYCLE

SAVE-FORCE SAVE-INDUCED SAVE-VELOCITY VERBOSE WRITEOUT

FORCE FIELD SELECTION KEYWORDS

FORCEFIELD PARAMETERS

POTENTIAL FUNCTION SELECTION KEYWORDS

ANGANGTERM ANGLETERM BONDTERM

CHARGETERM CHGDPLTERM DIPOLETERM

EXTRATERM IMPROPTERM IMPTORSTERM METALTERM MPOLETERM OPBENDTERM OPDISTTERM PITORSTERM POLARIZETERM

RESTRAINTERM RXNFIELDTERM SOLVATETERM STRBNDTERM STRTORTERM TORSIONTERM TORTORTERM UREYTERM VDWTERM

POTENTIAL FUNCTION PARAMETER KEYWORDS

ANGANG ANGLE ANGLE3

ANGLE4 ANGLE5 ANGLEF

ATOM BIOTYPE BOND

BOND3 BOND4 BOND5

CHARGE DIPOLE DIPOLE3

DIPOLE4 DIPOLE5 ELECTNEG

HBOND IMPROPER IMPTORS METAL MULTIPOLE OPBEND

OPDIST PIATOM PIBOND

PITORS POLARIZE SOLVATE STRBND STRTORS TORSION TORSION4 TORSION5 TORTOR

UREYBRAD VDW VDW14

VDWPR

ENERGY UNIT CONVERSION KEYWORDS

ANGLEUNIT ANGANGUNIT BONDUNIT

IMPROPUNIT IMPTORUNIT OPBENDUNIT OPDISTUNIT PITORSUNIT STRBNDUNIT

STRTORUNIT TORSIONUNIT TORTORUNIT UREYUNIT