Plate-based diversity subset screening generation 2: an improved paradigm for high-throughput screening of large compound files.

High-throughput screening (HTS) is an effective method for lead and probe discovery that is widely used in industry and academia to identify novel chemical matter and to initiate the drug discovery process. However, HTS can be time consuming and costly and the use of subsets as an efficient alternative to screening entire compound collections has been investigated. Subsets may be selected on the basis of chemical diversity, molecular properties, biological activity diversity or biological target focus. Previously, we described a novel form of subset screening: plate-based diversity subset (PBDS) screening, in which the screening subset is constructed by plate selection (rather than individual compound cherry-picking), using algorithms that select for compound quality and chemical diversity on a plate basis. In this paper, we describe a second-generation approach to the construction of an updated subset: PBDS2, using both plate and individual compound selection, that has an improved coverage of the chemical space of the screening file, whilst only selecting the same number of plates for screening. We describe the validation of PBDS2 and its successful use in hit and lead discovery. PBDS2 screening became the default mode of singleton (one compound per well) HTS for lead discovery in Pfizer.

Plate-based diversity subset screening: an efficient paradigm for high throughput screening of a large screening file.

The screening files of many large companies, including Pfizer, have grown considerably due to internal chemistry efforts, company mergers and acquisitions, external contracted synthesis, or compound purchase schemes. In order to screen the targets of interest in a cost-effective fashion, we devised an easy-to-assemble, plate-based diversity subset (PBDS) that represents almost the entire computed chemical space of the screening file whilst comprising only a fraction of the plates in the collection. In order to create this file, we developed new design principles for the quality assessment of screening plates: the Rule of 40 (Ro40) and a plate selection process that insured excellent coverage of both library chemistry and legacy chemistry space. This paper describes the rationale, design, construction, and performance of the PBDS, that has evolved into the standard paradigm for singleton (one compound per well) high-throughput screening in Pfizer since its introduction in 2006.

Shaping a screening file for maximal lead discovery efficiency and effectiveness: elimination of molecular redundancy.

High Throughput Screening (HTS) is a successful strategy for finding hits and leads that have the opportunity to be converted into drugs. In this paper we highlight novel computational methods used to select compounds to build a new screening file at Pfizer and the analytical methods we used to assess their quality. We also introduce the novel concept of molecular redundancy to help decide on the density of compounds required in any region of chemical space in order to be confident of running successful HTS campaigns.

Physiochemical drug properties associated with in vivo toxicological outcomes.

Relationships between physicochemical drug properties and toxicity were inferred from a data set consisting of animal in vivo toleration (IVT) studies on 245 preclinical Pfizer compounds; an increased likelihood of toxic events was found for less polar, more lipophilic compounds. This trend held across a wide range of types of toxicity and across a broad swath of chemical space.

Analysis of data fusion methods in virtual screening: theoretical model.

This paper presents a theoretical model of how data fusion can be used to combine the results of multiple similarity searches of chemical databases. The model is based on frequency distributions of similarity values that are fused using a multiple integration over regions defined by the particular fusion rule that is being applied. For pairwise fusion, the resulting double integrals are straightforward to evaluate for simple model distributions. Similarity values for recovered-active and recovered-nonactive frequency distributions are independently modeled using a constant background, linearly biased terms, and a first-order correlated term. The model shows that two standard fusion rules can give performance enhancements in some cases but that the results of fusion are dependent on many factors that, taken together, can lead to seemingly inconsistent levels of enhancement.

Analysis of data fusion methods in virtual screening: similarity and group fusion.

In a recent companion paper we have related the operation of simple data fusion rules used in virtual screening to a multiple integral formalism. In this paper we extend these ideas to the analysis of data fusion methods applied to real data. We examine several cases of similarity fusion using different coefficients and different representations and consider the reasons for positive or negative results in terms of the similarity distributions. Results are obtained using the SUM-, MAX- MIN-, and CombMNZ-fusion rules. We also develop a customized fusion rule, which provides an estimate of the optimal possible result for fusing multiple searches of a specific database; this shows that similarity fusion can, in principle, achieve retrieval enhancements even if this is not achieved in practice with current fusion rules. The methods are extended to analyze the comparatively successful results of group fusion with multiple actives, and we provide a rationale for the observed superiority of the MAX-rule over the SUM-rule in this context.

Enhancing the effectiveness of virtual screening by fusing nearest neighbor lists: a comparison of similarity coefficients.

This paper evaluates the effectiveness of various similarity coefficients for 2D similarity searching when multiple bioactive target structures are available. Similarity searches using several different activity classes within the MDL Drug Data Report and the Dictionary of Natural Products databases are performed using BCI 2D fingerprints. Using data fusion techniques to combine the resulting nearest neighbor lists we obtain group recall results which, in many cases, are a considerable improvement on standard average recall values obtained for individual structures. It is shown that the degree of improvement can be related to the structural diversity of the activity class that is searched for, the best results being found for the most diverse groups. The group recall of active compounds using subsets of the class is also investigated: for highly self-similar activity classes, the group recall improvement saturates well before the full activity class size is reached. A rough correlation is found between the relative improvement using the group recall and the square of the number of unique compounds available in all of the merged lists. The Tanimoto coefficient is found unambiguously to be the best coefficient to use for the recovery of active compounds using multiple targets. Furthermore, when using the Tanimoto coefficient, the "MAX" fusion rule is found to be more effective than the "SUM" rule for the combination of similarity searches from multiple targets. The use of group recall can lead to improved enrichment in database searches and virtual screening.