The present disclosure provides a method of generating compound information in a computing apparatus, the method including obtaining a learning model for information associated with partial structures, obtaining information associated with a source molecule that is a target of a partial structure modification, obtaining information associated with a partial structure set including a plurality of partial structures of the source molecule, selecting, from the partial structures included in the partial structure set, a target partial structure to be modified, obtaining, using the learning model, information associated with a modified partial structure corresponding to the target partial structure, and outputting result information in which the target partial structure is replaced by the modified partial structure in the source molecule.

The invention relates to an in silico screening method to identify candidate excipients for reducing aggregation of a protein in a formulation. The method combines computational molecular modeling and molecular dynamics simulations to identify sites on a protein where non-specific self-interaction and interaction of different test excipients may occur, determine the relative binding energies of such interactions, and select one or more test excipients that meet specified interaction criteria for use as candidate excipients in empirical screening studies.

The invention relates to an in silico screening method to identify candidate excipients for reducing aggregation of a protein in a formulation. The method combines computational molecular modeling and molecular dynamics simulations to identify sites on a protein where non-specific self-interaction and interaction of different test excipients may occur, determine the relative binding energies of such interactions, and select one or more test excipients that meet specified interaction criteria for use as candidate excipients in empirical screening studies.

Owing to complexity of the algorithms and tools very few attempts have been seen for usage of simulation methods in the development of new electrolytes. Moreover, the existing simulation methods focus on only one aspect of the electrolyte at a time and this limits accuracy of simulation results, and affects performance of electrolyte in real world, where multiple factors come into play simultaneously. The method disclosed provides method and system for in-silico optimization and design of electrolytes, enabling prediction of various properties of an electrolytic mixture of salts, solvents and various additives and its suitability for a given battery technology. The in-silico method shapes itself into an overall battery electrolyte property or component composition analyzer based on the user input.

Owing to complexity of the algorithms and tools very few attempts have been seen for usage of simulation methods in the development of new electrolytes. Moreover, the existing simulation methods focus on only one aspect of the electrolyte at a time and this limits accuracy of simulation results, and affects performance of electrolyte in real world, where multiple factors come into play simultaneously. The method disclosed provides method and system for in-silico optimization and design of electrolytes, enabling prediction of various properties of an electrolytic mixture of salts, solvents and various additives and its suitability for a given battery technology. The in-silico method shapes itself into an overall battery electrolyte property or component composition analyzer based on the user input.

This disclosure provides systems and methods for sample processing and data analysis. Sample processing may include nucleic acid sample processing and subsequent sequencing. Some or all of a nucleic acid sample may be sequenced to provide sequence information, which may be stored or otherwise maintained in an electronic storage location. The sequence information may be analyzed with the aid of a computer processor, and the analyzed sequence information may be stored in an electronic storage location that may include a pool or collection of sequence information and analyzed sequence information generated from the nucleic acid sample. Methods and systems of the present disclosure can be used, for example, for the analysis of a nucleic acid sample, for producing one or more libraries, and for producing biomedical reports. Methods and systems of the disclosure can aid in the diagnosis, monitoring, treatment, and prevention of one or more diseases and conditions.

This disclosure provides systems and methods for sample processing and data analysis. Sample processing may include nucleic acid sample processing and subsequent sequencing. Some or all of a nucleic acid sample may be sequenced to provide sequence information, which may be stored or otherwise maintained in an electronic storage location. The sequence information may be analyzed with the aid of a computer processor, and the analyzed sequence information may be stored in an electronic storage location that may include a pool or collection of sequence information and analyzed sequence information generated from the nucleic acid sample. Methods and systems of the present disclosure can be used, for example, for the analysis of a nucleic acid sample, for producing one or more libraries, and for producing biomedical reports. Methods and systems of the disclosure can aid in the diagnosis, monitoring, treatment, and prevention of one or more diseases and conditions.

A variational autoencoder (VAE) has been developed to learn a continuous numerical, or latent, representation of molecular structure to expand reference libraries for small molecule identification. The VAE has been extended to include a chemical property decoder, trained as a multitask network, to shape the latent representation such that it assembles according to desired chemical properties. The approach is unique in its application to metabolomics and small molecule identification, focused on properties that are obtained from experimental measurements (m/z, CCS) paired with its training paradigm, which involves a cascade of transfer learning iterations. First, molecular representation is learned from a large dataset of structures with m/z labels. Next, in silico property values are used to continue training. Finally, the network is further refined by being trained with the experimental data. The trained network is used to predict chemical properties directly from structure and generate candidate structures with desired chemical properties. The network is extensible to other training data and molecular representations, and for use with other analytical platforms, for both chemical property and feature prediction as well as molecular structure generation.

A variational autoencoder (VAE) has been developed to learn a continuous numerical, or latent, representation of molecular structure to expand reference libraries for small molecule identification. The VAE has been extended to include a chemical property decoder, trained as a multitask network, to shape the latent representation such that it assembles according to desired chemical properties. The approach is unique in its application to metabolomics and small molecule identification, focused on properties that are obtained from experimental measurements (m/z, CCS) paired with its training paradigm, which involves a cascade of transfer learning iterations. First, molecular representation is learned from a large dataset of structures with m/z labels. Next, in silico property values are used to continue training. Finally, the network is further refined by being trained with the experimental data. The trained network is used to predict chemical properties directly from structure and generate candidate structures with desired chemical properties. The network is extensible to other training data and molecular representations, and for use with other analytical platforms, for both chemical property and feature prediction as well as molecular structure generation.

Disclosed are methods for identifying bio-molecules with desired properties (or which are most suitable for a round of directed evolution) from complex bio-molecule libraries or sets of such libraries. Some embodiments of the present disclosure provide methods for virtually screening proteins for beneficial properties. Some embodiments of the present disclosure provide methods for virtually screening enzymes for desired activity and/or selectivity for catalytic reactions involving particular substrates. Some embodiments combine screening and directed evolution to design and develop proteins and enzymes having desired properties. Systems and computer program products implementing the methods are also provided.