Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for utilizing language machine learning model (LLM) as autonomous reasoners to navigate and execute multiple layers of a computerized bio-activity discovery pipeline of a tech-bio exploration system. In particular, the disclosed systems can utilize an LLM that learns to access one or more tech-bio exploration tools to execute one or more processes and/or tasks in a bio-activity discovery pipeline. For instance, the disclosed systems can provide an interactive query prompt interface to enable users to provide tech-bio queries (as prompts) and utilize the LLM with the prompts to execute one or more tasks in the bio-activity discovery pipeline to generate and/or retrieve bio-activity data for the query. Moreover, the disclosed systems can utilize one or more LLMs to autonomously utilize and/or interact with one or more tech-bio tools in the bio-activity discovery pipeline to generate and/or obtain bio-activity data.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that a implement a framework for active learning to discover pairwise interactions via representation learning. Indeed, in one or more implementations, the disclosed systems generate a first individual perturbation embedding from a first representation of a first cell exposed to a first perturbation and a second individual perturbation embedding, from a second representation of a second cell exposed to a second perturbation. For instance, the disclosed systems combine the first individual perturbation embedding and the second individual perturbation embedding to determine a predicted pairwise embedding. Moreover, in some instances, the disclosed systems generate a pairwise embedding from a representation of a cell exposed to both the first and second perturbation.

Abstract: In various embodiments, a molecule exploration application determines one or more potential drug candidates during a drug discovery process. The molecule exploration application generates derived molecule specifications based on a query molecule specification and edit heuristics. Subsequently, the molecule exploration application performs, via a mapping algorithm, one or more mapping operations on the derived molecule specifications to generate mapped molecule specifications. The molecule exploration application then performs one or more search operations on a mapped catalog of molecules based on the mapped molecule specifications to determine the one or more potential drug candidates. Advantageously, the molecule exploration application can be used to efficiently determine additional drug development candidates during a drug discovery process.

Abstract: One embodiment of the present invention sets forth a technique for training a machine learning model to generate image embeddings for images captured during multiple experiments. The technique includes inputting a batch of images into a plurality of layers in the machine learning model, wherein the batch of images has been sampled from a plurality of images generated via a first experiment. The technique also includes, for at least one layer included in the plurality of layers, computing a set of statistics associated with a plurality of outputs generated by the layer based on the batch of images and normalizing the plurality of outputs based on the statistics. The technique further includes updating a plurality of parameters for each of the plurality of layers based on a set of predictions generated by the first machine learning model from the batch of images and the normalized plurality of outputs.

Abstract: An aromatic heterocyclic compound, a pharmaceutical composition and use thereof. Specifically disclosed are a compound as shown in formula I, a stereoisomer thereof, a diastereomer thereof, or a pharmaceutically acceptable salt of any one of the foregoing, or a crystal form or solvate of any one of the foregoing. The aromatic heterocyclic compound has a novel structure, good CDK7 inhibitory activity, and good selectivity.

Abstract: The invention provides a compound of formula (I), or pharmaceutically acceptable ester, amide, carbamate, solvate or salt thereof, including a salt of such an ester, amide or carbamate, wherein R1 is an optionally substituted phenyl, or an optionally substituted 5- or 6-membered aromatic heterocycle; and R2 is an optionally substituted 5- or 6-membered aromatic heterocycle. Also provided are pharmaceutical compositions comprising a compound of formula (I).

Abstract: Provided herein are compounds of Formula (I) and Formula (II) and stereoisomers, tautomers, and pharmaceutically acceptable salts or solvates thereof. Also provided are pharmaceutical compositions comprising these compounds and methods of treating a disease, disorder, or condition by using the compounds and pharmaceutical compositions.

Abstract: A method for computational drug design includes defining a population of a plurality of compounds. Each compound includes one or more molecular properties. The method includes defining a training set of compounds from the population for which one or more biological properties are known. The method includes selecting, from the population, a subset of one or more compounds that are not in the training set. The method includes determining a subset score of the selected subset based on molecular properties of the one or more compounds in the selected subset, and evaluating the selected subset based on the determined subset score. The subset score is determined based on a frequency of the molecular properties in the population and on a frequency of the molecular properties in a sampled set comprising the training set and the selected subset.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that analyze gene perturbation machine learning embeddings and clinical observation data sets utilizing machine learning, explainability models, and causal discovery models to generate causal predictions between one or more genes and clinical outcomes. Indeed, in one or more implementations, the disclosed systems identify gene perturbation embeddings generated from cells exposed to perturbations. For instance, the disclosed systems select a cluster of genes from a plurality of genes by applying a clustering model to the gene perturbation embeddings. In some instances, the disclosed systems select gene targets from the cluster of genes by using a machine learning classification model trained on a plurality of features of the clinical observation data set.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that utilizing compound-protein machine learning representations to generate target results. For example, the disclosed systems can utilize a compound-protein interaction machine learning model to generate a compound-protein machine learning representation for compound protein pairs. The disclosed systems can utilize the compound-protein machine learning representation to train and utilize other target machine learning models in generating predicted bioactivity results. For example, the disclosed systems train a target machine learning model from compound-protein machine learning representations to generate ADMET predictions and/or biological perturbation program predictions.

Abstract: Systems and methods for determining whether a set of test perturbations discriminates over a null distribution for an on target effect against a first component of an entity are disclosed. The perturbations are perturbations of the first component and the entity comprises a plurality of components. For each perturbation in the set, a corresponding vector comprising a plurality of elements, is obtained. Each element comprises a distribution metric of measurements of a feature across instances of the entity upon exposure to the respective perturbation or (ii) a distribution metric of a respective dimension reduction component computed using the measurement of the plurality of features across instances of the entity upon the perturbation exposure. A composite metric is computed, using the vectors, and compared to a null distribution.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that utilizing compound-protein machine learning representations to generate target results. For example, the disclosed systems can utilize a compound-protein interaction machine learning model to generate a compound-protein machine learning representation for compound protein pairs. The disclosed systems can utilize the compound-protein machine learning representation to train and utilize other target machine learning models in generating predicted bioactivity results. For example, the disclosed systems train a target machine learning model from compound-protein machine learning representations to generate ADMET predictions and/or biological perturbation program predictions.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that implement an orchestration interface for managing and operating complex compound discovery pipelines across computer networks. Indeed, in one or more implementations, the disclosed systems manage and operate a compound discovery event publication repository for collecting and publishing events from various devices of different stages within the compound discovery pipeline according to a flexible event schema. To illustrate, utilizing the disclosed systems, individual devices can define and publish events with corresponding relationships (e.g., dependencies, parent-child relationships) for consumption and utilization by other group devices in a manner that provides flexibility to individual devices while coordinating accurately with downstream computer processes.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for embedding perturbation data via a machine learning model and filtering, aligning, and aggregating the embeddings to generate a genome-wide perturbation database for real-time generation of perturbation heatmaps. In particular, in one or more embodiments, the disclosed systems can receive a plurality of perturbation images portraying cells from a plurality of wells corresponding to a plurality of cell perturbations. Further, the systems can generate, utilizing a machine learning model, a plurality of well-level image embeddings from the plurality of perturbation images. Moreover, the systems can align, utilizing an alignment model, the plurality of well-level image embeddings to generate aligned well-level image embeddings. Additionally, the systems can aggregate, according to perturbations of one or more perturbation experiments, the well-level image embeddings to generate perturbation-level image embeddings.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that train and utilize machine learning models to generate perturbation embeddings from phenomic images of cells, including neuronal cell images. Indeed, in one or more implementations, the disclosed systems generate a perturbation embedding using an adapter model or a mixture of experts model. In some implementations, the disclosed systems utilize a mixture of experts model that combines phenomic embeddings from different embedding models to generate a mixture of experts phenomap that contains information from multiple embedding models.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods that implement a cross-platform flexible data model for dynamic storage, management, and retrieval of high-volume object data, including multi-modal machine learning datasets. Indeed, in one or more implementations, the disclosed systems operate across a variety of different non-tabular media and multiple different digital repository platforms, to efficiently and flexibly store, manage, and retrieve machine learning datasets by utilizing a dynamic cross-platform metadata database and cross-platform file location database. In particular, the disclosed systems can dynamically evolve metadata for data files of various machine learning datasets and independently manage storage locations across digital repository platforms by maintaining and utilizing a centralized cross-platform metadata database and corresponding cross-platform file location database.

Abstract: In a method of molecular scaffold hopping an interface of a scheduler computer sends instructions, prepared by the scheduler computer, to a job runner computer to perform a plurality of separate computational tasks. Each of the separate computational tasks includes calculating one or more chemical properties for a query molecule or molecules in a library of molecules. One or more of the plurality of separate computational tasks performed on the job runner computer are preemptible computing instances. Status indicators sent from the job runner computer are received by the interface for each of the plurality of separate computational tasks. The indicators are one of: incomplete, completed, or failed computing instances. The interface resends the instructions to the job runner computer that correspond to the separate computational tasks having the failed computing instance indicator to increase fault-tolerance against the separate computational tasks not attaining the completed computing instance indicator.

Abstract: Methods and systems for evaluating a query perturbation, in a cell based assay representing a test state, are provided. Control data points having dimensions representing measurements of different features across control cell aliquots are obtained. Test data points having dimensions representing measurements of different features across test cell aliquots are obtained. A composite test vector is computed between measures of central tendency across the control data points and measures of central tendency across the test data points. Query perturbation data points having dimensions representing measurements of different features across perturbation cell aliquots are obtained. A composite query perturbation vector is computed between measures of central tendency across the control data points and measures of central tendency across the plurality of query perturbation data points.

Abstract: Methods and systems for evaluating a query perturbation, in a cell based assay representing a test state, are provided. Control data points having dimensions representing measurements of different features across control cell aliquots are obtained. Test data points having dimensions representing measurements of different features across test cell aliquots are obtained. A composite test vector is computed between measures of central tendency across the control data points and measures of central tendency across the test data points. Query perturbation data points having dimensions representing measurements of different features across perturbation cell aliquots are obtained. A composite query perturbation vector is computed between measures of central tendency across the control data points and measures of central tendency across the plurality of query perturbation data points.

Abstract: The disclosure provides methods and systems for identifying a subset of compounds in a plurality of compounds. The identifying includes obtaining, for each compound, a vector including a set of elements, where each element includes a measurement of a different feature of an instance of a cell context upon exposure to the compound. The identifying includes performing the obtaining for a plurality of cell contexts, to obtain a plurality of vectors for each compound across different cell contexts. The identifying includes combining the vectors for each compound to form a combined vector for each compound, thereby forming a plurality of combined vectors representing different compounds. The identifying includes pruning the plurality of compounds to the subset of compounds based on a similarity between respective combined vectors in the plurality of combined vectors corresponding to compounds in the plurality of compounds.