This software calculates a large number of molecular structure         descriptors from the connection table of the submitted structure.  The descriptors         calculated fall into  several categories.            
     
              The outline below gives a listing of all the descriptors calculated by         the current  version.  The list is organized by index         sub-type.  Additional information is available by clicking on the topic of interest.            

              The first 8 columns of output from give information about the structure file         that has been submitted for each compound.  This information includes         a sequential ID from processing.      
     
        Several processing flags are also set by Molconn.  These are different         from the Molconn error messages that are discussed in section 5.  Processing         flags indicate that some change was made in the submitted structure before the         indices were calculated and the flags serve to notify the user of the change.            

              The Structure Information Representation (SIR) SUM indices may be used in conjunction         with the formula weight (fw) to identify duplicates in a data set.  The         proceedure for identifying duplicates is given at the bottom of this section.            

              Some elementary structure information indices are calculated for each         molecule that    is processed.  This information includes the formula weight and counts of         features such as rings, rotatable bonds and elements.            

      The       Electrotopological State or E-State structure indices are Structure-Information       representation descriptors that encode the electron accessibility of       atoms in a molecule.  This encodes both the electron distribution        and local topology at each atom in atom level and bond level indices.        The E-State is based on an intrinsic state for each atom that is       modified by all other atoms in the molecule.
     
      The atom and bond level indices are grouped in various ways to form       atom-type, hydrogen atom-type,  functional-group type, group-type, internal       hydrogen bonding and bond type indices.  Additional information about each       type of index, including a comprehensive list of available indices, may be viewed       by clicking on the corresponding link in the outline below.            

              The Atom-Level E-State, or "S", quantifies the build up or depletion of valence electrons         modified by the steric accessibility as the electron accessibility at the atom and         is calculated for each atom in the molecule.  In this form, the E-State         may be used as an atom-level descriptor.  The formalism is extended to         hydrogen atoms with the hydrogen atom level index "HS".  The HS index is         used most often with hydrogen atoms that have been explicitly enumerated  in         the structures.            

              This atom level description can be used to compare the electron accessibility         of    topologically equivalent positions within structures with a common         scaffold.           The example below shows two penicillin analogues.  The atom level         E-State values for atoms 1-16 may be used as structure descriptors because         these atom positions are         topologically equivalent in both structures.  The atom 5 position         can be used    despite the fact that structure B has no atom at the 5 position.  For         structure    B,    s5 will be set to s5 = 0.  The 17 and 18 positions         cannot    be    used    because    of the    ambiguity of the numbering at these         positions (more than one position in the structure could be numbered         17).            

              Global E-State indices have a value for every, or nearly         every, molecule processed and tend to be based on whole         molecule calculations.  For this reason, global         descriptors are often useful in providing broad classification         information such as lipophilicity and polarity.  Global         E-State descriptors tend to focus on the electronic characteristics         of a compound rather than size, shape and branching attributes which         are encoded in the molecular connectivity indices.            

              The Atom-Type E-State indices are an extension of the Atom-Level E-State,         similar to group additive schemes, in which an index value appears         for each atom type in the molecule.  The atom-type indices are based         on the valence state (SssO is all sp3 divalent oxygen         atoms) of the atom being described, irrespective of the functional groups that         the atom may be associated with.      
     
        The "Atom-Level" E-State discissed in section 5.4.1 quantifies the build up         or depletion of valence electrons modified by the steric accessibility as the electron         accessibility at the atom.  The "Atom-Type" E-State qualifies the kinds         of interactions that the specified electron accessibility may participate in.           Because atoms of a given type may be expected to engage in similar         interactions across a heterogenous group of structures, they provide         the basis for application to a wide range of problems in which the E-State         formalism is used without the need for atom-level superposition.            

              The Atom-Type E-State indices are an extension of the Atom-Level E-State,         similar to group additive schemes, in which an index value appears         for each atom type in the molecule.  Some atom types are based         on valence state (SssO is all divalent oxygen atoms) and some are based         on organic functional groups (SEther is all ether oxygen atoms).        
     
        The "Atom-Level" E-State discissed in section 5.3.1 quantifies the build up         or depletion of valence electrons modified by the steric accessibility as the electron         accessibility at the atom.  The "Atom-Type" E-State qualifies the kinds         of interactions that the specified electron accessibility may participate in.           Because these atom types and functional groups are expected to engage in similar         interactions across a heterogenous group of structures, they provide         the basis for application to a wide range of problems in which the E-State         formalism is used without the need for superposition.            

             Group types allow for combining the contributions of related atom        types or functional groups into a single broad-scope parameter, allowing        structure features to appear in a model in more than one form to reflect        multiple modes of interaction.  The use of Group-Types allows        for information about under-represented structure features to be included in a model        in cases where low population level might otherwise result in the        feature being overlooked by the feature selection algorithm.          Group-Types can be especially useful when modeling small datasets where the        total number of input parameters that can be used is limited by statistical concerns.            

             The E-State hydrogen bonding indices have been designed to encode the         potential for pairs of heteroatoms within a molecule to engage in internal        hydrogen bonding.  The index is constructed by calculating the         product of the atom-level E-State for the acceptor atom and the atom-level        hydrogen HE-State of the hydrogen atom on the donor.  SHBint        indices are enumerated for donor and acceptor paris that are separated        by  a given count of bonds.  The SHBint3, SHBint4, SHBint5 and        THBint3, THBint4, THBint5 indices represent atom pairs that can for stable        ring systems as a result of internal hydrogen bonds.  A screening        criteria has been applied to these pairs to remove sets where an internal        hydrogen bond would not be possible (non-compliant), such as donors and acceptors        that are meta substituted neighbors on a planar ring.      
     
       Although the SHBint2 and SHBint7-SHBint10 indices do not represent atom        pairs that can form internal hydrogen bonds, these indices have shown        up as important in studying properties such as "drug likeness".          The SHBint(n) indices give the largest product of E-State and HE-State        of the donor and acceptor pairs separated by n bonds.  The SHBI(n)        indices give the cumulative sum of E-State and HE-State products for all        donor        and acceptor pairs separated by n bonds.            

              The molecular connectivity and kappa shape indices provide information         about the kind of skeletal structure that a compound is based on.           The low order chi indices are related to molecular size, volume and surface         area.  The higher order chi indices encode information about skeletal         branching patterns, ring substitution and fused ring architecture.      

              The Bond-Type indices are an extension of the E-State formalism where a bond         is characterized using the two atoms that define the bond.  The intrinsic         state value for the bond is the geometric mean of the intrinsic         state values for the two atoms in the bond.  As in the atom-level         formalism, the actual bond E-State value is defined as the perturbed intrinsic         state of the bond, that is the intrinsic state plus the sum of perturbations         by all other bonds in the molecule.  The bond-level values are summed         for all identical bonds in the molecule to form the Bond-Type indices.            

              This page gives the list of Bond-Type E-State indices for bonds between         carbon atoms and other carbon atoms.            

              This page gives the list of Bond-Type E-State indices for bonds between         carbon atoms and nitrogen, onuim and onium pseudoion atoms.