 |
 |
 | FILTER
Database preparation Property calculation | ||||
 |
 |
 |
 |
||||
|
|
| FILTER
Database preparation Property calculation | ||||
|
|
|
|
||||
For this release, we have included three new experimental filters in addition to the default ``lead-like'' (filter_lead.txt) and ``drug-like'' (filter_drug.txt) filters. The filters filter_alldrug.txt, filter_blockbuster.txt and filter_2.5_blockbuster.txt can all be found in the data directory of the filter distribution. The filter_alldrug.txt filter is by far the most lenient filter available (though users are free to generate their own). This filter was designed to be the minimum filter that would pass all the molecules found in the ``small molecule drug'' file from the DrugBank web site (http://redpoll.pharmacy.ualberta.ca/drugbank) [15]. Unfortunately, this filter stands more as an example of the futility of this type of approach than as a practical filter. In order to pass every one of these molecules, including some that would not be acceptable for modern project work, many of the individual filters must be set to unreasonable values. For instance, the molecular weight range is 30 to 1500 and the hetero-atom count range is 1 to 60!
In order to generate more useful filters, rather than using a large list of drug molecules, we focussed on the best-selling, non-antibiotic, prescription drugs. We compiled a list of approximately 140 compounds. We designed two filters around this set of compounds, one that passes all of the compounds and another that is slightly more descriminating. For details of this second filter, please see ``Accumulation of rules'' above. Here we'll focus on the first and most useful of these filters, filter_blockbuster.txt. The physical property values in this filter are quite good. However, the functional group filters in this filter are probably too restrictive as 140 compounds is not sufficient to span all acceptable functionality. We expect that combining the physical property limits from filter_blockbuster.txt with the functional group filters from filter_lead.txt or filter_drug.txt may provide a very good filter. In future releases we plan to include a default filter that is optimized along these lines. For this release, that experimentation is left to the user. Despite the caveat concerning overly strict functional group filters, we expect that even without modification, filter_blockbuster.txt is a useful filter.
|
|
| FILTER
Database preparation Property calculation | ||||
|
|
|
|
||||