Volume 1 Chapter 6 2D Similarity Searching in Graphics QUEST3D

Back to Table of Contents

7.1 The 3D-CONSTRAIN Sub-menu

A 3D search normally involves the location of a chemical substructural fragment which has pre-defined 3D spatial characteristics. The chemical nature of the substructure is defined by use of the BUILD menu and 2D-CONSTRAIN sub-menu; the required 3D spatial characteristics are then defined using the 3D-CONSTRAIN sub-menu.

A typical, and simple, example of a 3D search is shown below. The bridged ring fragment is common to morphine and related alkaloids of pharmacological importance. Given that we know that the distance from the heterocyclic N atom to the centre of the aromatic ring is close to 4.6Å in parent morphine, then a search in which this distance is constrained to lie between 4.45 and 4.75Å should retrieve other fragments which have similar spatial characteristics.

QUEST3D

(a) 2D / 3D Query

(b) 2D / 3D Display of 'hit'

(c) Geometry table

 Nfrag  Refcod        NANG    N-C1    N-C5    DIST    PANG    NDEV
     4  BERCAC     112.566   1.501   1.482   4.464  89.456   1.647
     5  BIKZIE10   111.872   1.514   1.510   4.508  89.532   1.369
     6  BIKZOK10   111.590   1.521   1.505   4.460  85.208   1.549
    12  CAXPIA     111.669   1.514   1.514   4.487  86.537   1.496
    13  CBMBZM     111.975   1.527   1.486   4.470  87.259   1.610
    14  CEKWIY     111.548   1.506   1.469   4.522  86.371   1.166
    16  CIDTAK     111.971   1.463   1.452   4.643  75.939   0.419
    17  CIMMUG     108.868   1.551   1.495   4.502  88.763   1.240
    18  CIMNAN     112.019   1.522   1.515   4.602  78.437   0.646
    19  CIVGOD10   111.154   1.533   1.501   4.614  78.735   0.708
    20  CIVGUJ10   113.583   1.521   1.523   4.499  84.462   1.181

(d) Descriptive statistics

        Mean       112.163   1.503   1.491   4.526  85.063   1.160
        S.D.Sample   1.686   0.026   0.024   0.055   4.526   0.357
        S.D.Mean     0.216   0.003   0.003   0.007   0.580   0.046
        Minimum    107.089   1.442   1.445   4.451  74.791   0.419
        Maximum    117.741   1.565   1.574   4.643  89.820   1.698
        Nobs            61      61      61      61      61      61

GSTAT

(e) Histogram of DIST

DIST        2    4    6    8   10   12   14   16   18   20
       .....I....I....I....I....I....I....I....I....I.....
  4.45 -**                                               - (  1)
       .**************************                       . ( 13)
       .**************                                   . (  7)
       .**************                                   . (  7)
       .********                                         . (  4)
  4.52 -******                                           - (  3)
       .****                                             . (  2)
       .**************                                   . (  7)
       .********                                         . (  4)
       .********                                         . (  4)
  4.60 -******                                           - (  3)
       .********                                         . (  4)
       .**                                               . (  1)
       .**                                               . (  1)
       .                                                 .
  4.67 -                                                 -
       .                                                 .
       .                                                 .
       .                                                 .
       .                                                 .
  4.75 -                                                 -
       .....I....I....I....I....I....I....I....I....I.....
DIST        2    4    6    8   10   12   14   16   18   20

The chemical substructural fragment itself is encoded as a 2D search query as described in chapter 6 . The 3D geometrical descriptors and search limits are then added via the 3D-CONSTRAIN sub-menu. Of 93 fragments located in the 2D search, 61 have the N to ring-centroid distance in the required range. The distribution of these distances is shown in a histogram generated by GSTAT.

However, it is very important to realise that the word 'search' has a much broader meaning in the 3D case than it has in the case of text or 2D chemical connectivity information. The fact that we can place numerical constraints on a particular geometrical quantity implies that we have considerable knowledge of the 3D structural chemistry of a given fragment. Sadly, this is not generally the case.

For these reasons, the 'search' instructions form a small part of the 3D-CONSTRAIN sub-menu. An equally important function of this sub-menu is to permit variability in the choice of geometrical parameters used to characterise a fragment, and to allow the user to survey and examine distributions of these parameters over all instances of the fragment. These facilities are often used without the application of any search constraints at all, in order to establish the best set of descriptive parameters, and to indicate the most appropriate search limits.

In order to describe the meaning and arrangement of the buttons in the 3D-CONSTRAIN sub-menu, we must analyse the fundamental processes involved in a 3D search of the CSD. Indeed, it will be seen that the phrase "3D constraints" is an oversimplification of the facilities available through this sub-menu, which comprise:


Back to Table of Contents

Volume 1 Chapter 7 The Connectivity Basis of 3D Searching.