Comparison of 2D Fingerprint Methods for Multiple‐Template Similarity Searching on Compound Activity Classes of Increasing Structural Diversity

Abstract

We studied the similarity search performance of differently designed molecular fingerprints using multiple reference structures and different search strategies. For this purpose, nine compound activity classes were assembled that exclusively consisted of molecules with different core structures and that represented different levels of intra-class structural diversity. Thus, there was a strict one-to-one correspondence between test compounds and core structures. Analysis of unique core structures was found to be a better measure of class diversity than distributions of simplified scaffolds. On increasingly diverse classes, a trainable fingerprint using a unique search strategy performed better than others tested herein. Overall, clear preferences were detected for nearest-neighbor search strategies over fingerprint-averaging techniques. Nearest-neighbor searching that relied on selecting database compounds most similar to one of the reference structures often improved compound recovery over other averaging methods, but at the cost of decreasing the ability to detect hits that were structurally distinct from reference molecules.

Support-Vector-Machine-Based Ranking Significantly Improves the Effectiveness of Similarity Searching Using 2D Fingerprints and Multiple Reference Compounds

“…In addition, for SVM-based ranking, different percentages of 2D-ZINC were used as negative training examples: 0.001%, 0.01%, 0.1%, and 1% corresponding to 14, 144, 1442, and 14 423 compounds, respectively. Activity classes are abbreviated according to Table 1. print search calculations mirror trends that are generally observed; 29,30 that is, search performance is compound-classdependent and varies among different fingerprint designs. In the calculations reported here, on average, Molprint2D produced the highest recovery rates, followed by Daylight, TGT, TGD, and MACCS.…”

New strategy for Turbo Similarity Searching: Implementation and testing

“…The result obtained showed a preference of fingerprints to the nearest neighbours method, which is able to do so by using the multiple combination of information from the reference set as a whole [10].…”

“…Similar methods (i.e centroid and nearest neighbours) have been used by [10] with an addition of a newly introduced frequency method to perform multi-template searching using a range of 2D fingerprints. The result obtained showed a preference of fingerprints to the nearest neighbours method, which is able to do so by using the multiple combination of information from the reference set as a whole [10].…”

Advanced Fingerprint Methods for Similarity Searching: Balancing Molecular Complexity Effects

“…However, fingerprint searching usually becomes more effective when multiple reference compounds (and hence more chemical information) are available [2,24]. For fingerprint searching using multiple reference molecules, different methods have been introduced [2,10,11,24,25]. For example, fingerprint averaging -also known as the centroid method-or data fusion of multiple Tc values can be applied to compare a database compound to a reference set [7].…”