Use of connectivity mapping to support read across: A deeper dive using data from 186 chemicals, 19 cell lines and 2 case studies.

We previously demonstrated that the Connectivity Map (CMap) (Lamb et al., 2006) concept can be successfully applied to a predictive toxicology paradigm to generate meaningful MoA-based connections between chemicals (De Abrew et al., 2016). Here we expand both the chemical and biological (cell lines) domain for the method and demonstrate two applications, both in the area of read across. In the first application we demonstrate CMap's utility as a tool for testing biological relevance of source chemicals (analogs) during a chemistry led read across exercise. In the second application we demonstrate how CMap can be used to identify functionally relevant source chemicals (analogs) for a structure of interest (SOI)/target chemical with minimal knowledge of chemical structure. Finally, we highlight four factors: promiscuity of chemical, dose, cell line and timepoint as having significant impact on the output. We discuss the biological relevance of these four factors and incorporate them into a work flow.

Grouping 34 Chemicals Based on Mode of Action Using Connectivity Mapping.

Connectivity mapping is a method used in the pharmaceutical industry to find connections between small molecules, disease states, and genes. The concept can be applied to a predictive toxicology paradigm to find connections between chemicals, adverse events, and genes. In order to assess the applicability of the technique for predictive toxicology purposes, we performed gene array experiments on 34 different chemicals: bisphenol A, genistein, ethinyl-estradiol, tamoxifen, clofibrate, dehydorepiandrosterone, troglitazone, diethylhexyl phthalate, flutamide, trenbolone, phenobarbital, retinoic acid, thyroxine, 1α,25-dihydroxyvitamin D3, clobetasol, farnesol, chenodeoxycholic acid, progesterone, RU486, ketoconazole, valproic acid, desferrioxamine, amoxicillin, 6-aminonicotinamide, metformin, phenformin, methotrexate, vinblastine, ANIT (1-naphthyl isothiocyanate), griseofulvin, nicotine, imidacloprid, vorinostat, 2,3,7,8-tetrachloro-dibenzo-p-dioxin (TCDD) at the 6-, 24-, and 48-hour time points for 3 different concentrations in the 4 cell lines: MCF7, Ishikawa, HepaRG, and HepG2 GEO (super series accession no.: GSE69851). The 34 chemicals were grouped in to predefined mode of action (MOA)-based chemical classes based on current literature. Connectivity mapping was used to find linkages between each chemical and between chemical classes. Cell line-specific linkages were compared with each other and to test whether the method was platform and user independent, a similar analysis was performed against publicly available data. The study showed that the method can group chemicals based on MOAs and the inter-chemical class comparison alluded to connections between MOAs that were not predefined. Comparison to the publicly available data showed that the method is user and platform independent. The results provide an example of an alternate data analysis process for high-content data, beneficial for predictive toxicology, especially when grouping chemicals for read across purposes.

Tissue-specific profiling reveals transcriptome alterations in Arabidopsis mutants lacking morphological phenotypes.

Traditional genetic analysis relies on mutants with observable phenotypes. Mutants lacking visible abnormalities may nevertheless exhibit molecular differences useful for defining gene function. To examine this, we analyzed tissue-specific transcript profiles from Arabidopsis thaliana transcription factor gene mutants with known roles in root epidermis development, but lacking a single-gene mutant phenotype due to genetic redundancy. We discovered substantial transcriptional changes in each mutant, preferentially affecting root epidermal genes in a manner consistent with the known double mutant effects. Furthermore, comparing transcript profiles of single and double mutants, we observed remarkable variation in the sensitivity of target genes to the loss of one or both paralogous genes, including preferential effects on specific branches of the epidermal gene network, likely reflecting the pathways of paralog subfunctionalization during evolution. In addition, we analyzed the root epidermal transcriptome of the transparent testa glabra2 mutant to clarify its role in the network. These findings provide insight into the molecular basis of genetic redundancy and duplicate gene diversification at the level of a specific gene regulatory network, and they demonstrate the usefulness of tissue-specific transcript profiling to define gene function in mutants lacking informative visible changes in phenotype.

Smart pooling of mRNA samples for efficient transcript profiling.

Gene expression profiling studies are commonly used to study signaling pathways and their impact on transcriptional regulation in plants. In some cases, a profiling study results in expression profiles in which most genes exhibit a small number of differentially expressed states among a large number of samples. In such instances, a pooling approach would help improve the efficiency of the profiling effort by employing fewer microarray chips and ensuring more robust measurement of transcript levels. Smart pooling involves pooling of mRNA samples in an information-efficient manner such that each sample is tested multiple times but always in pools with other samples. The resulting pooled measurements are then decoded to recover the expression profile of all samples in the study. In this protocol, we describe in detail the process of designing smart pooling experiments and decoding their results, which have been used for studying signaling in Arabidopsis root development. Heuristics are provided to select the design parameters that would ensure successful execution of smart pooling.

A gene regulatory network for root epidermis cell differentiation in Arabidopsis.

The root epidermis of Arabidopsis provides an exceptional model for studying the molecular basis of cell fate and differentiation. To obtain a systems-level view of root epidermal cell differentiation, we used a genome-wide transcriptome approach to define and organize a large set of genes into a transcriptional regulatory network. Using cell fate mutants that produce only one of the two epidermal cell types, together with fluorescence-activated cell-sorting to preferentially analyze the root epidermis transcriptome, we identified 1,582 genes differentially expressed in the root-hair or non-hair cell types, including a set of 208 "core" root epidermal genes. The organization of the core genes into a network was accomplished by using 17 distinct root epidermis mutants and 2 hormone treatments to perturb the system and assess the effects on each gene's transcript accumulation. In addition, temporal gene expression information from a developmental time series dataset and predicted gene associations derived from a Bayesian modeling approach were used to aid the positioning of genes within the network. Further, a detailed functional analysis of likely bHLH regulatory genes within the network, including MYC1, bHLH54, bHLH66, and bHLH82, showed that three distinct subfamilies of bHLH proteins participate in root epidermis development in a stage-specific manner. The integration of genetic, genomic, and computational analyses provides a new view of the composition, architecture, and logic of the root epidermal transcriptional network, and it demonstrates the utility of a comprehensive systems approach for dissecting a complex regulatory network.

poolMC: smart pooling of mRNA samples in microarray experiments.

BACKGROUND: Typically, pooling of mRNA samples in microarray experiments implies mixing mRNA from several biological-replicate samples before hybridization onto a microarray chip. Here we describe an alternative smart pooling strategy in which different samples, not necessarily biological replicates, are pooled in an information theoretic efficient way. Further, each sample is tested on multiple chips, but always in pools made up of different samples. The end goal is to exploit the compressibility of microarray data to reduce the number of chips used and increase the robustness to noise in measurements. RESULTS: A theoretical framework to perform smart pooling of mRNA samples in microarray experiments was established and the software implementation of the pooling and decoding algorithms was developed in MATLAB. A proof-of-concept smart pooled experiment was performed using validated biological samples on commercially available gene chips. Differential-expression analysis of the smart pooled data was performed and compared against the unpooled control experiment. CONCLUSIONS: The theoretical developments and experimental demonstration in this paper provide a useful starting point to investigate smart pooling of mRNA samples in microarray experiments. Although the smart pooled experiment did not compare favorably with the control, the experiment highlighted important conditions for the successful implementation of smart pooling - linearity of measurements, sparsity in data, and large experiment size.

Pooling in high-throughput drug screening.

Pooling in HTS refers to the act of testing mixtures of compounds in a primary screen to accurately identify hits for secondary screening. The reduction in the number of tests needed to screen a compound library by pooling can also be extended to achieve much-needed error tolerance in HTS. Despite the success of HTS in other biological experiments, pooling in high-throughput drug screening has been a controversial and often marginalized paradigm. At first appearance, pooling appears to promise gains from reduced effort, or possibly could create more problems than solutions. However, this article demonstrates that pooling is a practical and necessary part of HTS: discussions include the rationale for pooling compounds in HTS, a unifying view of pooling design theory, a review of past attempts at pooling and their success, and recent advances in the field.

poolHiTS: a shifted transversal design based pooling strategy for high-throughput drug screening.

BACKGROUND: A key goal of drug discovery is to increase the throughput of small molecule screens without sacrificing screening accuracy. High-throughput screening (HTS) in drug discovery involves testing a large number of compounds in a biological assay to identify active compounds. Normally, molecules from a large compound library are tested individually to identify the activity of each molecule. Usually a small number of compounds are found to be active, however the presence of false positive and negative testing errors suggests that this one-drug one-assay screening strategy can be significantly improved. Pooling designs are testing schemes that test mixtures of compounds in each assay, thereby generating a screen of the whole compound library in fewer tests. By repeatedly testing compounds in different combinations, pooling designs also allow for error-correction. These pooled designs, for specific experiment parameters, can be simply and efficiently created using the Shifted Transversal Design (STD) pooling algorithm. However, drug screening contains a number of key constraints that require specific modifications if this pooling approach is to be useful for practical screen designs. RESULTS: In this paper, we introduce a pooling strategy called poolHiTS (Pooled High-Throughput Screening) which is based on the STD algorithm. In poolHiTS, we implement a limit on the number of compounds that can be mixed in a single assay. In addition, we show that the STD-based pooling strategy is limited in the error-correction that it can achieve. Due to the mixing constraint, we show that it is more efficient to split a large library into smaller blocks of compounds, which are then tested using an optimized strategy repeated for each block. We package the optimal block selection algorithm into poolHiTS. The MATLAB codes for the poolHiTS algorithm and the corresponding decoding strategy are also provided. CONCLUSION: We have produced a practical version of STD algorithm for pooled drug screens. This pooling strategy provides both assay compression and error-correction capabilities that can both accelerate and reduce the overall cost of HTS in drug discovery.