Data were vegas casino download not available for all the selected compounds in all available historical screens, and we set a threshold of a minimum number of 50 active and 50 inactive datapoints required for a screen to be included. Corresponding activity data were extracted from an internal HTS assay database. The compounds set was selected based on chemical diversity, known annotations, compound availability, and on general representation in historical HTS screens.

Comparing bioactivity predictions using different input modalities

A set of 8,300 compounds was selected and screened in the Cell Painting assay. In a prior study by Simm et al., 342 top-ranked compounds were validated for a kinase target in oncology and 141 top-ranked compounds were validated for a non-kinase enzyme in a CNS indication. Employing bootstrapping, we consistently observed enrichment ranging from a moderate 1.6-fold to a high 14-fold, correlating with the predictive performance of the assays. We sampled and tested at least 1,000 compounds for each assay, encompassing the majority of the top-ranked 5% of compounds, along with at least 500 compounds selected uniformly at random. Nonetheless, noticeable performance variations exist among different assay and target types, as illustrated in Fig. Our approach could significantly reduce the size of screening campaigns, saving time and resources, and enabling primary screening with more complex assays or scarce material.

Our state-of-the-art platforms employ a wide array of detection technologies enabling superior sensitivity in screening and detection of invaluable hits with high confidence. The company has invested in 10 fully automated HTS platforms including BSL2+/3-enabled platforms for high throughput screening of microorganisms, ensuring rapid generation of robust data. Our team of experts work with you to design the most suitable screening strategy for your target or phenotype. Discover a treasure trove of tractable chemical starting points in Evotec’s extensive collection of compounds, carefully curated to fuel groundbreaking discoveries. This collection provides access to diverse and novel biologically relevant chemical space, and the compounds have been selected for their chemical tractability and drug-likeness. High throughput screening (HTS) is one of the most widely used methods for hit identification.

• With an excellent combination of disease biology knowledge as well as drug discovery and development expertise, Evotec selects the most suitable screening strategy for your target class and therapeutic area of interest.
• In fact, two out of the four follow-up assays performed slightly better than their respective primary assays (Fig. 4a and Supplementary Table 1).
• The general trend we found was that the predictive performance was consistently good for different assay types.
• Overall, the Cell Painting fluorescence-based approach performed best both in terms of correctly predicting bioactivity (measured by ROC-AUC) and in terms of increased chemical diversity (measured by Tanimoto similarity).
• Initially, we assessed our framework’s performance on a dataset established by Hofmarcher and colleagues8, demonstrating end-to-end learning with convolutional neural networks (CNNs) for biological assay prediction from Cell painting images.

Evotec’s Compound Collection

The colors -red, white, and blue in the heatmap represent the relative expression  levels as high, medium, and low respectively. 106c.f.u per mL were transferred in a 96-well  microtiter plate and the primary screening was performed in the epMOTION 5075. The dose-response curve  may inform us about the efficacy of a library compound or an unknown small molecule. If  screening is performed for the inhibitory effect, then positive controls would consist of a  molecule that is lethal to the bacterium of interest. When designing experiments for small molecule screening, appropriate controls are required.

We demonstrate efficient learning despite the inherent noise in single-point activity data. In addition, the previous approaches evaluated assays at 3–4 thresholds of pXC50/IC50/EC50. In these cases, each activity label is often supported by multiple data points (usually 10–20), significantly enhancing the confidence in the labels. The other assays, Oxidoreductase, Polymerase, and Methyltransferase, all had follow-up enrichments in line with the primary assay (6.4x, 4.6x, and 1.6x respectively). A Receiver operating curves for the four assays, primary HTS (green) and follow-up (blue), shaded area represents standard-deviation interval. In fact, two out of the four follow-up assays performed slightly better than their respective primary assays (Fig. 4a and Supplementary Table 1).

How To Optimize Your Hit Identification Strategy – Evotec

Recently there has been a strong interest in combining compound structure information with activity fingerprints leading to improved performance in bioactivity prediction4. ROC-AUC values for each of the four assays were calculated using the randomly sampled subset of compounds, green showing the average performance in the original HTS assays and blue representing the performance when using secondary screen activity readouts. We put the predictions of our Cell Painting-based model to the test by running secondary assays for the same targets. Previous studies that have aimed to predict bioactivity from image-based assays have limited their analysis to a single primary assay. 1b, the predictive performance of the Cell Painting fluorescence-image based model varied widely from assay to assay, ranging from 0.96 to 0.48 ROC-AUC.

High Throughput Screening (HTS) Services

The authors would like to thank Guy Williams, Diane Smith, Hannah Semple, and Elizabeth Mouchet for their invaluable help in producing the Cell painting dataset used in this publication. The publicly available Cell Painting data used in this study are available from the JUMP consortium dataset13, CPG0016 available from the Cell Painting Gallery on the Registry of Open Data on AWS. The raw HTS datasets generated and analysed in this study are protected and are not available due to them being AstraZeneca proprietary information.

Before being sent to the network as input during training, the images were augmented, including spatial down sampling, z-normalization, random cropping, horizontal and vertical flipping, random 90-degree rotations and color shifting. The images were pre-processed and normalized such that the top and bottom 1 percentile intensity values were clipped for each image to remove noise and outliers. Fluorescent microscopy images were stored as 16-bit TIFFs of size 1992×1992. We report both the mean ROC-AUC over all assays as well as the individual ones.

As it has been shown that brightfield images can be used to predict Cell Painting features15, we wanted to investigate if the information content in the images would be sufficient to predict bioactivity of compounds. To further strengthen the practicality of our approach, we analyze prediction performance and robustness across various assay types, technologies, and target classes to identify specific targets and assays that are particularly well-suited for bioactivity prediction. Furthermore, we explore different input modalities for bioactivity prediction, including fluorescence images, brightfield images, and image features extracted from the fluorescence images using classical image analysis approaches. We employ a large-scale general purpose Cell Painting screen to capture phenotypic profiles of a library of available compounds and train a model using small, focused bioactivity assay readouts for specific targets. So-called structure activity relation (SAR) models are a family of computational methods, used to make bioactivity predictions or property predictions i.e., using computational methods and models to estimate bioactivity or properties of chemical compounds. Top ranked compounds in four of the assays were selected for follow-up validation in secondary screening.

Cell based High Throughput Screening Assays of Bacteria.

We found the ROC-AUC values in the follow-up assays to be consistent with the values from the primary assays. Among molecular target subtypes, kinase targets appeared to benefit the most from our Cell Painting-based approach, performing significantly better than other molecular target subtypes (Fig. 3d). The general trend we found was that the predictive performance was consistently good for different assay types. Therefore, we conducted a detailed analysis, breaking down the results to examine how various assay characteristics contribute to performance (Fig. 3).

We investigate the potential of deep learning on unrefined single-concentration activity readouts and Cell Painting data, to predict compound activity across 140 diverse assays. HTS uses miniaturised assays and automation to screen large compound libraries therefore generating data rapidly and cost effectively. HTS identifies potential hits which show binding or activity against a particular biological target or cellular phenotype. It involves screening thousands or millions of compounds through a previously developed biological assay using automation. Assay development is one of the most critical stages in the hit identification process as the quality of the assay and robustness of the automation infrastructure determines the quality of the data. Our capabilities include target-based approaches (e.g., biochemical, cell-based, biophysical, including ASMS and DEL technology, fragment-based screening etc), various cellular and phenotypic screening and in silico approaches.

The serine kinase assay, which had the highest predictive performance (ROC-AUC 0.91) showed an astoundingly high enrichment of 14x in the follow-up, representing a significant improvement and suggests this assay could focus on a small, highly targeted set of compounds. A The predictive performance of the image-based Fluorescence model compared to the structure based, when grouped by Test Material Type. For each bioactivity prediction approach, we compared the structural diversity of the 20 top-ranked compounds to the known actives in the training set. This subset encompassed 29 assays comprising 10,660 unique compounds (See Materials and Methods 1.3. JUMP consortium and ChEMBL datasets for details).

Our analysis revealed that compounds predicted from images showed lower structural similarity i.e., greater chemical diversity, than structure-based approaches. Although the brightfield image-based approach was outperformed by the fluorescence-based approach, it was still able to predict 49% of the assays with a ROC-AUC above 0.7 and even 5% above 0.9. This dataset included 209 assays comprising 10,574 compounds8,12, where binary activity data was derived from dose-response curves (IC50/EC50) of each compound in a given assay. Initially, we assessed our framework’s performance on a dataset established by Hofmarcher and colleagues8, demonstrating end-to-end learning with convolutional neural networks (CNNs) for biological assay prediction from Cell painting images. Our results demonstrate the capability of models trained on phenotypic data combined with a few hundred single-concentration data points, to predict compound activity reliably and efficiently across diverse targets in a realistic drug screening scenario. As only a few hundred activity data points are needed to train the predictive model for a particular target and assay, assays of higher complexity and biological relevance could potentially be used.

• Using a cross-validation setup, we split the data into 6 different folds with each compound only included in one-fold.
• Evaluation on the held-out test sets revealed that the predictive performance of the whole image fluorescence-based approach outperformed all other approaches (Fig. 2a).
• Because the initial screening assays are often very simple representations of the target biology, they run the risk of producing false positive and negative results.
• We investigate the potential of deep learning on unrefined single-concentration activity readouts and Cell Painting data, to predict compound activity across 140 diverse assays.

While the predictive performance of brightfield images does not fully reach the level of fluorescence images, brightfield offers a cheaper alternative for phenotypic profiling. Our approach has the potential to reduce the number of compounds screened, as well as the number of assays and experiments required in drug screening cascades, which in turn could allow for early screening of focused compound sets in assays of higher biological relevance. The results showed that Cell Painting-based bioactivity prediction using morphological profiles was feasible for a wide range of targets. This indicates that the model’s predictions are driven by the targets and phenotypes, and not significantly affected by biases and noise in the assays.

Moreover, several compounds were predicted to be active by the model and were confirmed to be active, despite them having been labeled as inactive in the original HTS data. B Enrichment values of the top 5% predicted compounds for each of the four assays. Among the therapy areas covered in our experiments, performance in oncology assays was significantly better than other areas, possibly due to a higher fraction of kinase targets in that therapy area. The increased performance on kinase targets could be attributed to the known promiscuity of kinase inhibitors, which can affect multiple cellular pathways, leading to stronger phenotypic responses that are more readily identified by the model.

Our results show that a model trained using phenotypic data from a single general-purpose Cell Painting screen can predict bioactivity in a wide range of assays, outperforming commonly used SAR models in terms of both predictive performance and structure diversity. Top ranked compounds suggested from the fluorescence image-based models were screened in corresponding secondary assays (see text for full description). In combination with these data, we rely on Cell Painting images of the full set of compounds – an additional cost – but one which only needs to be produced once and can then be reused across all the different assays and targets we want to predict (Fig. 1a).

Because the initial screening assays are often very simple representations of the target biology, they run the risk of producing false positive and negative results. Because of this, hit finding is generally done with simple assays such as biochemical assays to enrich the compound set before more resource-intense assays can be used further down the cascade. Accurate bioactivity prediction using morphological profiles could streamline the process, enabling smaller, more focused compound screens. Another important aspect of cell-based HT assays is the response of the organism of interest through the primary screen. On the other hand, cell-based assays discussed include viability, reporter gene, second messenger, and high-throughput microscopy assays.

The Brightfield image-based model was trained following the same procedure as for the fluorescent image-based model, although the default setting of three channels as input was used, stacking the three focal planes into one image. The identified hyper-parameters were then used to train the pre-trained ResNet50 for 100 epochs, using early stopping based on the validation ROC-AUC performance and learning rate stopping on plateau. The linear layer of the pre-trained model was replaced with a re-initialized one with 140 output neurons to match the number of assays.

Triple-effect correction for Cell Painting data with contrastive and domain-adversarial learning

No data were excluded from the analysis and the investigators were not blinded. Searching for optimal, model-depth number of hidden layers, layer-width, weight-decay and learning rate. Binary-cross entropy combined with focal loss was used to train the model. Using RDKit36, all compound SMILES40 representations were converted to ECFP4. Searching for optimal, model-depth, layer-width, weight-decay, learning rate and optimizer. Like the previous two model types, binary-cross entropy combined with focal loss was used to train the model.