Benchmarking feature selection methods for compressing image information in high-content screening.

Biopharmaceutical drug discovery, as of today is a highly automated, high throughput endeavor, where many screening technologies produce a high-dimensional measurement per sample. A striking example is High Content Screening (HCS), which utilizes automated microscopy to systematically access the wealth of information contained in biological assays. Exploiting HCS to its full potential traditionally requires extracting a high number of features from the images to capture as much information as possible, then performing algorithmic analysis and complex data visualization in order to render this high-dimensional data into an interpretable and instructive information for guiding drug development. In this process, automated feature selection methods condense the feature set to reduce non-useful or redundant information and render it more meaningful. We compare 12 state-of-the-art feature selection methods (both supervised and unsupervised) by systematically testing them on two HCS datasets from drug screening imaging assays of high practical relevance. Considering as evaluation metrics standard plate-, assay- or compound statistics on the final results, we assess the generalizability and importance of the selected features by use of automated machine learning (AutoML) to achieve an unbiased evaluation across methods. Results provide practical guidance on experiment design, optimal sizing of a reduced feature set and choice of feature selection method, both in situations where useful experimental control states are available (enabling use of supervised algorithms) or where such controls are unavailable, using unsupervised techniques.

Deep Learning-Based HCS Image Analysis for the Enterprise.

Drug discovery programs are moving increasingly toward phenotypic imaging assays to model disease-relevant pathways and phenotypes in vitro. These assays offer richer information than target-optimized assays by investigating multiple cellular pathways simultaneously and producing multiplexed readouts. However, extracting the desired information from complex image data poses significant challenges, preventing broad adoption of more sophisticated phenotypic assays. Deep learning-based image analysis can address these challenges by reducing the effort required to analyze large volumes of complex image data at a quality and speed adequate for routine phenotypic screening in pharmaceutical research. However, while general purpose deep learning frameworks are readily available, they are not readily applicable to images from automated microscopy. During the past 3 years, we have optimized deep learning networks for this type of data and validated the approach across diverse assays with several industry partners. From this work, we have extracted five essential design principles that we believe should guide deep learning-based analysis of high-content images and multiparameter data: (1) insightful data representation, (2) automation of training, (3) multilevel quality control, (4) knowledge embedding and transfer to new assays, and (5) enterprise integration. We report a new deep learning-based software that embodies these principles, Genedata Imagence, which allows screening scientists to reliably detect stable endpoints for primary drug response, assess toxicity and safety-relevant effects, and discover new phenotypes and compound classes. Furthermore, we show how the software retains expert knowledge from its training on a particular assay and successfully reapplies it to different, novel assays in an automated fashion.

Polar transmembrane-based amino acids in presenilin 1 are involved in endoplasmic reticulum localization, Pen2 protein binding, and γ-secretase complex stabilization.

γ-Secretase is composed of the four membrane proteins presenilin, nicastrin, Pen2, and Aph1. These four proteins assemble in a coordinated and regulated manner into a high molecular weight complex. The subunits constitute a total of 19 transmembrane domains (TMD), with many carrying important amino acids involved in catalytic activity, interaction with other subunits, or in ER retention/retrieval of unassembled subunits. We here focus on TMD4 of presenilin 1 (PS1) and show that a number of polar amino acids are critical for γ-secretase assembly and function. An asparagine, a threonine, and an aspartate form a polar interface important for endoplasmic reticulum retention/retrieval. A single asparagine in TMD4 of PS1 is involved in a protein-protein interaction by binding to another asparagine in Pen2. Intriguingly, a charged aspartate in TMD4 is critical for γ-secretase activity, most likely by stabilizing the newly formed complex.

Masking of transmembrane-based retention signals controls ER export of gamma-secretase.

gamma-Secretase is critically involved in the Notch pathway and in Alzheimer's disease. The four subunits of gamma-secretase assemble in the endoplasmic reticulum (ER) and unassembled subunits are retained/retrieved to the ER by specific signals. We here describe a novel ER-retention/retrieval signal in the transmembrane domain (TMD) 4 of presenilin 1, a subunit of gamma-secretase. TMD4 also is essential for complex formation, conferring a dual role for this domain. Likewise, TMD1 of Pen2 is bifunctional as well. It carries an ER-retention/retrieval signal and is important for complex assembly by binding to TMD4. The two TMDs directly interact with each other and mask their respective ER-retention/retrieval signals, allowing surface transport of reporter proteins. Our data suggest a model how assembly of Pen2 into the nascent gamma-secretase complex could mask TMD-based ER-retention/retrieval signals to allow plasma membrane transport of fully assembled gamma-secretase.

Endoplasmic reticulum retention of the gamma-secretase complex component Pen2 by Rer1.

gamma-Secretase is involved in the production of amyloid beta-peptide, which is the principal component of amyloid plaques in the brains of patients with Alzheimer disease. gamma-Secretase is a complex composed of presenilin (PS), nicastrin, anterior pharynx-defective phenotype 1 (Aph1) and PS enhancer 2 (Pen2). We previously proposed a mechanism of complex assembly by which unassembled subunits are retained in the endoplasmic reticulum (ER) and only the fully assembled complex is exported from the ER. We have now identified Retention in endoplasmic reticulum 1 (Rer1) as a protein that is involved in the retention/retrieval of unassembled Pen2 to the ER. Direct binding of unassembled Pen2 to Rer1 is mediated by the first transmembrane domain of Pen2, and a conserved asparagine in this domain is required. Downregulation of Rer1 leads to increased surface localization of Pen2, whereas overexpression of Rer1 stabilizes unassembled Pen2. To our knowledge, Rer1 is the first identified interaction partner of mammalian transmembrane-based retention/retrieval signals.