UNITY descriptors

In addition to looking for data trends in physical property space using PCA and PLS, trends in chemical structure space can be delineated by viewing nonlinear maps (NLM) of two-dimensional structure descriptors such as Unity Fingerprints or topological atom pairs using tools such as Benchware DataMiner [42]. Two-dimensional NLM plots provide an overview of chemical structure space and biological activity/molecular properties are mapped in a 3rd and/or 4th dimension to look for trends in the dataset. 

In contrast to partitioning methods that involve dimension reduction of chemical reference spaces, MP is best understood as a direct space method. However, -dimensional descriptor space is simplified here by transforming property descriptors with continuous or discrete value ranges into a binary classification scheme. Essentially, this binary space transformation assigns less complex -dimensional vectors to test molecules, with each dimension having unity length of either 0 or 1. Thus, although MP analysis proceeds in -dimensional descriptor space, its dimensions are scaled and its complexity is reduced. 

Autoscaling involves standardizing the measurement variables so that each descriptor or measurement has a mean of zero and a standard deviation of unity, that is,... 

After a brief summary of the molecular and MO-communication systems and their entropy/information descriptors in OCT (Section 2) the mutually decoupled, localized chemical bonds in simple hydrides will be qualitatively examined in Section 3, in order to establish the input probability requirements, which properly account for the nonbonding status of the lone-pair electrons and the mutually decoupled (noncommunicating, closed) character of these localized a bonds. It will be argued that each such subsystem defines the separate (externally closed) communication channel, which requires the individual, unity-normalized probability distribution of the input signal. This calls for the variable-input revision of the original and fixed-input formulation of OCT, which will be presented in Section 4. This extension will be shown to be capable of the continuous description of the orbital(s) decoupling limit, when AO subspace does not mix with (exhibit no communications with) the remaining basis functions. 

The suitability of these descriptors for database mining was investigated in a comparative analysis (Cruciani et al. 2002). While 2D descriptors like Unity... 

Matter [45] has also validated a range of 2D and 3D structural descriptors for their ability to predict biological activity and for their ability to be able to sample structurally and biologically diverse datasets effectively. The descriptors examined included Unity 2D fingerprints [42], atom-pairs [47],... 

Substructure searching is often used in drug design and needs no further clarification. Similarity searching is also a very well known technique described in more detail elsewhere [52], We usually use M ACCS keys, Unity fingerprints, CATS descriptors, and feature trees for similarity searching [53], Each technique has its own strengths and weaknesses, so we favor parallel application of two or three of them. 

One way to try to alleviate the problem of correlated descriptors is to perform a principal components analysis (see Section 9.13). Those principal components which explain (say) 90% of the variance may be retained for the subsequent calculations Alternatively, those principal components for which the associated eigenvalue exceeds unity may be chosen, or the principal components may be selected using more complex approaches based on cross-validation (see Section 12.12.3). It may be important to scale the descriptors (e.g. using autoscaling) prior to calculating the principal components. However, unless each principal component is largely associated with any particular descriptor it can be difficult to interpret the physical meaning of any subsequent results. ... 

The rules in Table 3.10 are characterized by the following parameters, which provide a more detailed account of the rule characteristics than provided in the earlier example described in Sections 3.2.3-3.2.S. Left-Hand-Side Support (LHS Supp) is the number of instances of the antecedent (i.e., IF part ) of a rule, while Right-Hand-Side Support (RHS Supp) is the number of instances of the consequent (i.e., THEN part ) of a rule. Since the rules are deterministic, these values should be the same, and hence, RHS Accuracy (RHS Acc), which is the ratio of LHS to RHS Support, and is equal to unity for aU of the cases considered here. LHS Coverage (LHS Cov) is the ratio of LHS Supp to the number of total number of objects (drugs), while RHS Coverage (RHS Cov) is the ratio of the RHS Supp to the number of objects (drugs) associated with the descriptor (phospholipidosis, O). The rules are... 

---