Search for Structurally Similar Molecules

In our view, this (now more than 15 years old) approach represents a significant advance in structure-activity studies that ought to have greater visibility in the SAR and QS AR community. Instead, using the words of Lahana and cowoikers in outlining their work, we prefer to cite the complete abstract of their publication as this will best describes the novelty of their approach [1]  

With this, we would like to end the introduction into justification for searching for novel molecular descriptors specifically designed for searching combinatorial libraries and highly similar structures. We hope that this may alert researchers looking for novel drugs to discuss this problem and, despite a few additional applications. 

one may consider construction of new molecular descriptors of very high discriminatory power specifically designed for screening combinatorial libraries one unsolved problem in this area. Observe that Lahana and coworkers used available molecular descriptors, most of which have limited discrimination powers and have been intended for use in MRA. There have been few molecular descriptors of very high discrimination power available, such as the molecular ID numbers [2,3-5], but they may not be suitable for screening combinatorial libraries because they are computationally intensive. 

When the row sums as entries in construction of the connectivity-type indices are applied to the adjacency matrix of molecular graphs, one obtains the connectivity index % of Randid [10,11]. When they are applied to the graph distance matrix, one obtains Balaban s index J [12]. Both indices, x and J, have been very nseful in QSAR, which is an incentive to explore properties of additional structural indices of this type. At our disposal are several novel sparse matrices that have been recently introduced in chemical graph theory, which, in the view that their row sums cover a wider range of values, can be expected to produce novel highly discriminatory topological indices. 

The Hamming distance between these two binary sequences is 7, which is the number of columns in which the two sequences (of equal length) have different entries. 

---