Introduction

The calculation of molecular properties from structure has always been closely associated with the Daylight and its progenitors. For many, the calculation of CLOGP was their first introduction to our software. With the advent of virtual screening there has been a resurgence in the requirements for rapid estimation of appropriate properties directly from structure. To that end, Daylight is introducing a Property Package which will provide values for most of the commonly used properties. The architecture of the package allows for additional properties to be added should they be required. Use of the program object technology ensures that the same code ( and hence the same answers ) are accessible from many programs including DayCart™

Program Objects

Program objects are used to provide two-way communication with an external process, e.g., the clogptalk program for computing hydrophobicity for a structure represented in SMILES. Using program objects, a calling program can start an external program, send it input, receive its output, and perform other tasks while the external program remains running and ready for more input. This is particularly important for many property programs which spend significant amounts of time initializing themselves.

A number of programs supporting program objects are supplied with the release of Daylight Software. Most of these are supplied as contributed code, in the directory: $DY_ROOT/contrib/src/c/progob/. The commercial programs clogptalk and cmrtalk operate as program objects (in $DY_ROOT/bin). The propertytalk program will operate in the same way, using the PIPETALK protocol.

Using program objects with Daycart™

The PIPETALK protocol provides an ideal way for the Daylight Oracle cartridge to communicate with external function for instance to populate a table column with clogp values say with a sql line like

UPDATE my_table SET clogP = clogp(smiles);

The propertytalk program

One of the advantages of talk programs is that you only need to have one copy of the code and a whole variety of programs can access the same algorithm/code. This means, for physical property calculations for instance, for a given structure you always get the same answer, irrespective of the calling program. To this end we have produced a talk program which will calculate a variety of physical properties commonly used in molecular selection and design directly from the structures. These values can then be stored along with the structure from which they were derived. These functions are available via a program_object interface which allows them to be called from within DayCart™. described above. Many of these are included in the VCS building routines, calculated from the parent or version smiles, as appropriate. All values are returned as part_tuples except for PART_COUNT which is an integer, this ensures that data for mixtures and salts are handled correctly.

1. Counting the number of matches of a pattern of atoms and bonds, usually expressed as SMARTS.
2. Additive constitutive properties where the patterns are counted and multiplied by a weighting factor e.g. molecular weight
3. Daylight object properties such as atom, cycle or part counts
4. Results of a Daylight algorithm such as fingerprint, depiction, conformation etc

In the future we will allow users to alter the paterns which are matched and the weights which are applied to the individual counts.

• AVERAGE_MOL_WT: Reads SMILES, calculates molecular weight based on average atomic weights for naturally occurring elements.
   
• MOL_FORM: Reads SMILES, calculates molecular formula in Hill order. Charges are ignored.
   
• ROT_BONDS: Reads SMILES, calculates count of rotatable bonds, secondary nitrogen connected to a trigonal carbon is fixed. I.e.
  [!$(*#*)&!D1]-&!@[!$(*#*)&!D1] rotates,
  [N&H1&D2]-&!@[#6&X3] does not.
   
• H_DONOR: Reads SMILES, calculates count of hydrogen bond donors.
  Donors are [!#6;!H0] .
  Note the number of donatable H's are counted not the heavy atoms to which the are attached.
   
• H_ACCEPTOR: Reads SMILES, calculates count of hydrogen bond acceptors.
  Acceptors are [$([!#6;+0]);!$([F,Cl,Br,I]);!$([o,s,nX3]);!$([Nv5,Pv5,Sv4,Sv6])].
  Note the number of acceptor sites are counted not the heavy atoms.
   
• PARACHOR: Reads SMILES, calculates parachor using McGowan's method, only works for C H N O S P F Cl Br I.
   
• CLOGP: Call commercial version.
   
• LOGP_STAR: Call commercial version.
   
• CMR: Call commercial version.
   
• ACCURATE_MASS: Reads SMILES, calculates molecular weight based on the accurate mass (IUPAC 1989 ) for most abundant isotope of each element.
  These values will be replaced with IUPAC 1999 values in the release version.
   
• MOLAR_VOLUME: Reads SMILES, calculates the average molar volume based on Schrödinger's method, only works for C H N O S F Cl Br I.
   
• RING_COUNT: Reads SMILES, calculates the count of rings in the SSSR ie the smallest set of smallest rings.
   
• ATOM_COUNT: Reads SMILES, calculates the count of heavy atoms (i.e. non-H ) in the molecule. Used to moderate the molecular weight values.
   
• RIGIDITY: Reads SMILES, calculates the Tanimoto similarity between the input molecule and the set of molecules formed by removing the rotatable bonds.A value of 1.00 implies the structure is totally rigid. Highly flexible molecules have much lower values, nearer zero .
   
• FRAG_COUNT: Reads SMILES, calculates the count of the number of fragments formed by removal of the isolating carbons from the structure.
  Isolating carbons are [$([#6]);!$(C(F)(F)F);!$(c(:[!c]):[!c]);!$([#6]=,#[!#6]); !$([#6;!+0])].
   
• FLEXIBILITY: Reads SMILES, calculates the flexibility of the structure from the ratio of the count of rotatable bonds, to the total count of bonds.
  Totally flexible structures have a value near 1.0.
  Rigid structures return a value of zero.
   
• FINGERPRINT: Reads SMILES, returns the lexical string version of the DAYLIGHT fingerprint. Default parameters are size 2048, min/max path 0/7.
   
• DEPICTION: Reads SMILES, returns a set of pairs of coordinates, in input SMILES order. Only coordinates for stereo significant hydrogens are returned.
   
• STEREO_COUNT: Reads SMILES, returns the number of potential stereo centres. These are necessary but not sufficient conditions for stereo. Atom stereo [$([X4&!v6&!v5;H0,H1]),$([SX3]([#6])([#6])~O)]
  Bond stereo [CX3;!H2]=[CX3;!H2]
  Allene stereo [CX3;H0]=C=[CX3;H0,H1]
   
• PART_COUNT: Reads SMILES, returns the number of components in a molecule.
   
• POLAR_SURFACE_AREA: Reads SMILES, returns the topological polar surface area, according to the method of P. Ertl, B. Rohde, P. Selzer "Fast Calculation of Molecular Polar Surface Area as a Sum of Fragment-based Contributions and Its Application to the Prediction of Drug Transport Properties", J.Med.Chem.(2000), 43, 371 4- 3717.
   
• USER_PATTERN: Reads SMILES, returns the number of unique matches of a user defined SMARTS in a molecule.
   

These programs can be accessed in the a demo form with the caveat that as it is running in a stateless world the program object is not persistent.

Plans and Futures

This package will be made available with a selection of some or all of the above properties in a 4.8 release. The appropriate SQL extensions are also available for use in conjunction with Daycart™. These may be accessed either in a parallel fashion or in a single user mode, depending on system resources.

In a later version there will be the equivalent of the algorithm manager in CLOGP and users will be allowed to edit the patterns which are mapped for the properties. This may be either via the talk protocol or an external file.

Daylight Chemical Information Systems, Inc. support@daylight.com