Robust Hit Identification by Quality Assurance and Multivariate Data Analysis of a High-Content, Cell-Based Assay

Dürr, Oliver; Duval, Frédéric; Nichols, Anthony; Lang, Paul; Brodte, Annette; Heyse, Stephan; Besson, Dominique

doi:10.1177/1087057107309036

Cited by 36 publications

(33 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For reducing the number of parameters describing the cell images, supervised methods have been used and reviewed in the literature. 10,25,26 Such methods select parameters based on whether a given set of control samples can be classified correctly. This limits the use of these methods to cases for which control samples are available and when one seeks to identify compounds exactly leading to the same phenotype as displayed in the positive controls.…”

Section: Discussionmentioning

confidence: 99%

“…8,9 Nevertheless, there are many examples in the literature demonstrating the benefits of a multiparametric analysis of HCS data. For example, the validation rate of hits determined using a support vector machine (SVM) was higher than using a standard analysis, 10 or a compound's target could be predicted based on a phenotypic profile. 11 It was also shown that by integrating phenotypic profiles obtained by HCS with chemical similarity, mode of action (MoA) could be inferred.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Comparison of Multivariate Data Analysis Strategies for High-Content Screening

et al. 2011

View full text Add to dashboard Cite

Section: Discussionmentioning

confidence: 99%

mentioning

confidence: 99%

Comparison of Multivariate Data Analysis Strategies for High-Content Screening

et al. 2011

View full text Add to dashboard Cite

“…Alternatively, supervised machine learning techniques such as linear discriminant analysis, support vector machines, or neural networks could be employed. 7 Nevertheless, the choice of a suitable training set for these methods can be challenging since overfitting of the models toward the phenotypes that are part of the training set can occur.…”

Section: Comparison Of Similarity Measure Performancementioning

confidence: 99%

“…It has been shown that hit identification with multiple HCS readouts can lead to lower rates of false positives and therefore reduced attrition rates at the hit verification stage. 6,7 Image-based readouts are composed of morphological, geometric, intensity, and texture-based features that can be combined to create a mathematical vector. This feature vector represents the treatment-induced phenotypic effect and provides a biologically relevant descriptor of a compound, which can be regarded as a signature or fingerprint and is thus named HCS fingerprint here.…”

Section: Introductionmentioning

confidence: 99%

Benchmarking of Multivariate Similarity Measures for High-Content Screening Fingerprints in Phenotypic Drug Discovery

et al. 2013

View full text Add to dashboard Cite

High-content screening (HCS) is a powerful tool for drug discovery being capable of measuring cellular responses to chemical disturbance in a high-throughput manner. HCS provides an image-based readout of cellular phenotypes, including features such as shape, intensity, or texture in a highly multiplexed and quantitative manner. The corresponding feature vectors can be used to characterize phenotypes and are thus defined as HCS fingerprints. Systematic analyses of HCS fingerprints allow for objective computational comparisons of cellular responses. Such comparisons therefore facilitate the detection of different compounds with different phenotypic outcomes from high-throughput HCS campaigns. Feature selection methods and similarity measures, as a basis for phenotype identification and clustering, are critical for the quality of such computational analyses. We systematically evaluated 16 different similarity measures in combination with linear and nonlinear feature selection methods for their potential to capture biologically relevant image features. Nonlinear correlation-based similarity measures such as Kendall's τ and Spearman's ρ perform well in most evaluation scenarios, outperforming other frequently used metrics (such as the Euclidian distance). We also present four novel modifications of the connectivity map similarity that surpass the original version, in our experiments. This study provides a basis for generic phenotypic analysis in future HCS campaigns.

show abstract

“…HCS is the cell-based quantification of several processes simultaneously, which provides a more detailed representation of the cellular response to various perturbations compared to HTS. HCS has many advantages over HTS 18,19 , but conducting a high-throughput (HT)-high-content (HC) screen in neuronal models is problematic due to high cost, environmental variation and human error. In order to detect cellular responses on a 'phenomics' scale using HC imaging one has to reduce variation and error, while increasing sensitivity and reproducibility.…”

mentioning

confidence: 99%

High Content Screening in Neurodegenerative Diseases

Jain¹,

Kesteren²,

Heutink³

2012

JoVE

View full text Add to dashboard Cite

The functional annotation of genomes, construction of molecular networks and novel drug target identification, are important challenges that need to be addressed as a matter of great urgency [1][2][3][4] . Multiple complementary 'omics' approaches have provided clues as to the genetic risk factors and pathogenic mechanisms underlying numerous neurodegenerative diseases, but most findings still require functional validation 5. For example, a recent genome wide association study for Parkinson's Disease (PD), identified many new loci as risk factors for the disease, but the underlying causative variant(s) or pathogenic mechanism is not known 6,7 . As each associated region can contain several genes, the functional evaluation of each of the genes on phenotypes associated with the disease, using traditional cell biology techniques would take too long.There is also a need to understand the molecular networks that link genetic mutations to the phenotypes they cause. It is expected that disease phenotypes are the result of multiple interactions that have been disrupted. Reconstruction of these networks using traditional molecular methods would be time consuming. Moreover, network predictions from independent studies of individual components, the reductionism approach, will probably underestimate the network complexity 8 . This underestimation could, in part, explain the low success rate of drug approval due to undesirable or toxic side effects. Gaining a network perspective of disease related pathways using HT/HC cellular screening approaches, and identifying key nodes within these pathways, could lead to the identification of targets that are more suited for therapeutic intervention.High-throughput screening (HTS) is an ideal methodology to address these issues [9][10][11][12] . but traditional methods were one dimensional whole-well cell assays, that used simplistic readouts for complex biological processes. They were unable to simultaneously quantify the many phenotypes observed in neurodegenerative diseases such as axonal transport deficits or alterations in morphology properties 13,14 . This approach could not be used to investigate the dynamic nature of cellular processes or pathogenic events that occur in a subset of cells. To quantify such features one has to move to multi-dimensional phenotypes termed high-content screening (HCS) 4,[15][16][17] . HCS is the cell-based quantification of several processes simultaneously, which provides a more detailed representation of the cellular response to various perturbations compared to HTS. HCS has many advantages over HTS 18,19 , but conducting a high-throughput (HT)-high-content (HC) screen in neuronal models is problematic due to high cost, environmental variation and human error. In order to detect cellular responses on a 'phenomics' scale using HC imaging one has to reduce variation and error, while increasing sensitivity and reproducibility.Herein we describe a method to accurately and reliably conduct shRNA screens using automated cell culturing 20 and H...

show abstract

Robust Hit Identification by Quality Assurance and Multivariate Data Analysis of a High-Content, Cell-Based Assay

Cited by 36 publications

References 15 publications

Comparison of Multivariate Data Analysis Strategies for High-Content Screening

Comparison of Multivariate Data Analysis Strategies for High-Content Screening

Benchmarking of Multivariate Similarity Measures for High-Content Screening Fingerprints in Phenotypic Drug Discovery

High Content Screening in Neurodegenerative Diseases

Contact Info

Product

Resources

About