A method includes generating a graph of a chemical compound based on at least one of an order-dependent representation of the chemical compound and a molecular graph representation of the chemical compound, encoding the graph based on an adjacency matrix of a graph convolutional neural network (GCN), an activation function of the GCN, and one or more weights of the GCN to generate a latent vector representation of the chemical compound, and decoding the latent vector representation based on a plurality of hidden states of a neural network (NN) to generate a reproduced order-dependent representation of the chemical compound.

In some embodiments, the early feasibility assessment (EFA) system generates a set of quantitative systems pharmacology (QSP) models using rule-based reasoning systems to efficiently assess the feasibility of therapeutic drug candidates. With each QSP model, the EFA system provides sets of parameters for the user to easily contrast and compare various scenarios to explore risk and uncertainty of developing the therapeutic drug candidate, while high-performance computing provides near real-time simulation of the scenarios. The EFA system performs 1-dimensional or 2-dimensional scans of parameters and output feasibility criteria including, for example, maximum inhibition, activation, and target engagement. The assessment results and key parameters values can be presented via a user interface of the EFA system which allows user interactions with, for example, dose-response and pharmacokinetic and pharmacodynamic (PK/PD) plots.

In some embodiments, the early feasibility assessment (EFA) system generates a set of quantitative systems pharmacology (QSP) models using rule-based reasoning systems to efficiently assess the feasibility of therapeutic drug candidates. With each QSP model, the EFA system provides sets of parameters for the user to easily contrast and compare various scenarios to explore risk and uncertainty of developing the therapeutic drug candidate, while high-performance computing provides near real-time simulation of the scenarios. The EFA system performs 1-dimensional or 2-dimensional scans of parameters and output feasibility criteria including, for example, maximum inhibition, activation, and target engagement. The assessment results and key parameters values can be presented via a user interface of the EFA system which allows user interactions with, for example, dose-response and pharmacokinetic and pharmacodynamic (PK/PD) plots.

Exemplary embodiments provide computer-implemented methods, mediums, and apparatuses configured to upload data stored in a data storage ecosystem to a cloud-based storage service. A database in the data storage ecosystem may store results sets from an analytical chemistry system. The results sets may be stored in a first model structure implemented by a library structure. An uploader may incorporate the library structure and may include logic to use the library structure to transform the results sets from first model structure into a second model structure suitable for use in a relational data store in the cloud-based storage system. By implementing the library structure in the uploader, the uploader can be decoupled from the data storage ecosystem. This allows the uploader to function without some of the overhead used by the data ecosystem, provide faster data uploads, and to automatically generate derived information for the results sets.

Exemplary embodiments provide computer-implemented methods, mediums, and apparatuses configured to upload data stored in a data storage ecosystem to a cloud-based storage service. A database in the data storage ecosystem may store results sets from an analytical chemistry system. The results sets may be stored in a first model structure implemented by a library structure. An uploader may incorporate the library structure and may include logic to use the library structure to transform the results sets from first model structure into a second model structure suitable for use in a relational data store in the cloud-based storage system. By implementing the library structure in the uploader, the uploader can be decoupled from the data storage ecosystem. This allows the uploader to function without some of the overhead used by the data ecosystem, provide faster data uploads, and to automatically generate derived information for the results sets.

A non-transitory computer-readable recording medium stores a compound substitution program for causing a computer to execute processing including: specifying a first partial structure included in a first compound; referring to information that indicates a relationship between a plurality of partial structures and selecting a second partial structure related to the first partial structure; specifying a bonding position in the second partial structure, based on a rational formula of the selected second partial structure; and generating information that indicates a second compound obtained by substituting the first partial structure of the first compound with the second partial structure, based on the specified bonding position.

A non-transitory computer-readable recording medium stores a compound substitution program for causing a computer to execute processing including: specifying a first partial structure included in a first compound; referring to information that indicates a relationship between a plurality of partial structures and selecting a second partial structure related to the first partial structure; specifying a bonding position in the second partial structure, based on a rational formula of the selected second partial structure; and generating information that indicates a second compound obtained by substituting the first partial structure of the first compound with the second partial structure, based on the specified bonding position.

Disclosed are systems and methods for generating candidate recipes for batch testing battery recipes in robotics laboratory equipment. In one embodiment, the candidate recipes in a batch, share the maximum number of chemicals in common, while as a batch, they utilize a minimum number of chemicals. The candidate recipes are identified by constructing a graph where an initial selection of recipes are placed at each node. The graph yields the candidate recipes in the batch.

Disclosed are systems and methods for generating candidate recipes for batch testing battery recipes in robotics laboratory equipment. In one embodiment, the candidate recipes in a batch, share the maximum number of chemicals in common, while as a batch, they utilize a minimum number of chemicals. The candidate recipes are identified by constructing a graph where an initial selection of recipes are placed at each node. The graph yields the candidate recipes in the batch.

A graph-to-sequence (G2S) architecture is configured to use graph data of objects to generate sequence data of new objects. The process can be used with objects types that can be represented as graph data and sequence data. For instance, such data is molecular data, where each molecule can be represented as molecular graph and in SMILES. Examples also include popular tasks in deep learning of image-to-text or/and image-to-speech translations. Images can be naturally represented as graphs, while text and speech can be natively represented as sequences. The G2S architecture can include a graph encoder and sample generator that produce latent data in a latent space, which latent data can be conditioned with properties of the object. The latent data is input into a discriminator to obtain real or fake objects, and input into a decoder for generating the sequence data of the new objects.