1.15.4. Saving pKa Results

Predicted molecular properties are ordinarily saved during a batch run as a tab-delimited data (.dat) file. This format is unsuitable for saving detail-rich results of the S+pKa model: variable number of macroscopic pKa, variable number of microstates, microconstants, pK50, H dissociation probabilities, etc. However, other output formats governed by option –u include this extra information. Chemical database software, such as popular in the past ISIS/Base™, provides a mechanism called hierarchical tables for storing a variable number of fields per record in a tree-like structure. Command line option –z was specifically designed with such hierarchical output in mind: it saves all S+pKa results including all microstates -u in an RDF file directly importable by ISIS/Base. It is likely that newer chemical database software also understands the RDF format and is capable of correctly placing all the S+pKa details for each molecule. Please consult appropriate IT specialist in your company regarding this aspect.

While option –z follows the “all-or-nothing” principle, option –w saves S+pKa results under much more selective regime. The key selection criteria here are “workflows” – output patterns meeting specific needs of different professional groups: medicinal chemists, physical chemists, cheminformaticists, and computational chemists. These workflows have been designed at Bayer Healthcare AG as part of a collaboration focused on expanding applicability domain of our S+pKa model onto the chemical space more characteristic of the pharmaceutical industry. Some of the output quantities (e.g., charge state) are pH-dependent where the pH is supplied with option –p.

The following results are saved for each kind of workflow:

Medicinal Chemist (one record per molecule):

Chemical structure in its neutral form
User input data
Predicted apparent pKa (SimPlus lowest acidic pK_a (strongest acid)) and prevalent group (SimPlus most acidic atom number(group)) with its pK50 value (SimPlus lowest acidic pK50) for the strongest acid (if present)
Predicted apparent pKa (SimPlus 2nd lowest acidic pK_a (strongest acid)) and prevalent group (SimPlus 2nd most acidic atom number(group)) with its pK50 value (SimPlus 2nd lowest acidic pK50) for the second strongest acid (if present)
Predicted apparent pKa (SimPlus 3rd lowest acidic pKa (strongest acid)) and prevalent group (SimPlus 3rd most acidic atom number(group)) with its pK50 value (SimPlus 3rd lowest acidic pK50) for the third strongest acid (if present)
Predicted apparent pK_a (SimPlus highest basic pKa (strongest base)) and prevalent group (SimPlus most basic atom number(group)) with its pK50 value (SimPlus highest basic pK50) for the strongest base (if present)
Predicted apparent pK_a (SimPlus 2nd highest basic pK_a (strongest base)) and prevalent group (SimPlus 2nd most basic atom number(group)) with its pK50 value (SimPlus 2nd highest basic pK50) for the second strongest base (if present)
Predicted apparent pKa (SimPlus 3rd highest basic pKa (strongest base)) and prevalent group (SimPlus 3rd most basic atom number(group)) with its pK50 value (SimPlus 3rd highest basic pK50) for the third strongest base (if present)
Predominant charge state (SimPlus predominant charge state at pH) at pH given by option –p (or 7.4 if –p is not present): neutral, positively charged, negatively charged, or zwitterionic
Numeric code for the charge state (SimPlus CSno) at pH defined above: 1 = negatively charged, 2 = zwitterionic, 3 = neutral, 4 = positively charged. A value of 0 that no charge was calculated, typically because there were too many ionizable centers..
Indicator of a successful prediction (SimPlus Success): 1 = success, 0 = failure
Full list of pK50 values for all ionizable groups (SimPlus acidic pK50s and SimPlus basic pK50s) = ionizable atom numbers followed by their pK50 value.

Group numbers in 2-5 and charge states in 6 are followed by quantitative measures of prevalence expressed in %.

Physical Chemist (multiple records per molecule, used to save individual microstates independent of pH):

Structures of ionization microstates in their respective charge forms
User input data
Predicted apparent pK_a (SimPlus macroscopic pKa) for the respective macrostate the saved microstate belongs to
Overall charge state (SimPlus overall charge state) (as numerical sum of all formal charges) of the saved microstate
Microstate prevalence in % (SimPlus microstate prevalence [%]) with respect to its parent macrostate, independent of pH
Indicator of a successful prediction (SimPlus Success): 1 = success, 0 = failure

Option –s filters out the output of minor microstates, or allows them all, if threshold is zero.

Cheminformaticist (one record per molecule):

Chemical structure in its neutral form
User input data
Charge state prevalence in % (SimPlus charge state prevalence [%] at pH) at pH given by option –p (or 7.4 if –p is not present)
Predominant charge state (SimPlus predominant charge state at pH) at pH defined above: neutral, positively charged, negatively charged, or zwitterionic
Numeric code for the charge state (SimPlus CSno) at pH defined above: 1 = negatively charged, 2 = zwitterionic, 3 = neutral, 4 = positively charged
Indicator of a successful prediction (SimPlus Success): 1 = success, 0 = failure

Computational Chemist (multiple records per molecule, used to save individual microstates dominant at a given pH):

Structures of ionization microstates in their respective charge forms
User input data
Predicted apparent pK_a (SimPlus macroscopic pK_a) for the respective macrostate the saved microstate belongs to
Overall charge state (SimPlus overall charge state) (as numerical sum of all formal charges) of the saved microstate
Microstate abundance (SimPlus microstate prevalence [%] at pH) in water solution, in %, at pH given by option –p (or 7.4 if –p is not present)
Indicator of a successful prediction (SimPlus Success): 1 = success, 0 = failure

Option –s filters out the output of minor microstates