A computer-implemented method of designing a molecule and determining a route to synthesise the molecule is provided. The method comprises: receiving one or more desired properties of the molecule; generating one or more candidate molecules using a first machine learning technique that uses the one or more desired properties of the molecule as an input; and for at least one candidate molecule, computing one or more routes to synthesise the candidate molecule using a second machine learning technique.

Described in certain example embodiments herein are systems, methods, and uses thereof for high-throughput in vitro evaluating multiple test compounds in parallel for biological or pharmacological functions. In certain embodiments, the system allows the selection of a subset of test compounds from a group of test compounds to form an optimized pool, and methods are provided to use such optimized pool of test compounds to identify and validate therapeutic agents for treating diseases and driving guided differentiation of stem cells into desired types of cells. The systems described herein can provide, for example, a cost-effective and high-quality high-throughput approach for drug screening.

Described in certain example embodiments herein are systems, methods, and uses thereof for high-throughput in vitro evaluating multiple test compounds in parallel for biological or pharmacological functions. In certain embodiments, the system allows the selection of a subset of test compounds from a group of test compounds to form an optimized pool, and methods are provided to use such optimized pool of test compounds to identify and validate therapeutic agents for treating diseases and driving guided differentiation of stem cells into desired types of cells. The systems described herein can provide, for example, a cost-effective and high-quality high-throughput approach for drug screening.

A computer implemented method for designing chemical reactions for catalyst construction is described. The method includes extracting historical data including historic chemical reaction data and historic catalyst construction yield data and converting the historic chemical reaction data into graph models to represent molecular structure data. The method also includes incorporating the graph models into a chemical reaction algorithm and training a vectorized cognitive deep learning network of the chemical reaction algorithm by using the graph models and a property of the historic chemical reaction data to produce a catalyst chemical reaction model. Further, the method includes validating the catalyst chemical reaction model by inputting the historic chemical reaction data and comparing a generated property corresponding to the catalyst chemical reaction model to the property of the historic chemical reaction data. Lastly, the method includes updating the training of the catalyst chemical reaction model.

A computer implemented method for designing chemical reactions for catalyst construction is described. The method includes extracting historical data including historic chemical reaction data and historic catalyst construction yield data and converting the historic chemical reaction data into graph models to represent molecular structure data. The method also includes incorporating the graph models into a chemical reaction algorithm and training a vectorized cognitive deep learning network of the chemical reaction algorithm by using the graph models and a property of the historic chemical reaction data to produce a catalyst chemical reaction model. Further, the method includes validating the catalyst chemical reaction model by inputting the historic chemical reaction data and comparing a generated property corresponding to the catalyst chemical reaction model to the property of the historic chemical reaction data. Lastly, the method includes updating the training of the catalyst chemical reaction model.

A computational model may be used to predict targets of a candidate, or predict candidates that interact with a target. A plurality of pairs may be established, each including a candidate and a respective one of a plurality of controls, each of the plurality of controls known to bind with a target. For each pair, values of at least two datatypes of the candidate may be compared to values of the at least two datatypes of the respective one of the plurality of controls in the pair to generate a similarity score for each of the at least two datatypes of each pair. Similarity scores may be converted to likelihood values indicating likelihood that the candidate and the controls have a shared target based on the respective one of the at least two datatypes. Tests may be performed to validate predictions regarding interactivity of candidates and targets.

A computational model may be used to predict targets of a candidate, or predict candidates that interact with a target. A plurality of pairs may be established, each including a candidate and a respective one of a plurality of controls, each of the plurality of controls known to bind with a target. For each pair, values of at least two datatypes of the candidate may be compared to values of the at least two datatypes of the respective one of the plurality of controls in the pair to generate a similarity score for each of the at least two datatypes of each pair. Similarity scores may be converted to likelihood values indicating likelihood that the candidate and the controls have a shared target based on the respective one of the at least two datatypes. Tests may be performed to validate predictions regarding interactivity of candidates and targets.

The system (200) for training an ensemble neural network device configured to execute the steps of: providing (205) a set of exemplar data, comprising at least one set of inputs (220) and at least one set of outputs (225) associated to the set of inputs, to a neural network device comprising an ensemble (230) of neural network devices, configured to provide independent predictions based upon the exemplar data, operating (210) the neural network device based upon the set of exemplar data, obtaining (215) the trained neural network device configured to provide an output, the neural network device further comprising at least two independent activation functions, whereof at least two of the independent activation functions are representative of the statistical distribution of the plurality of independent predictions, the neural network device being configured to provide at least one output (235, 236) for at least two said independent activation functions and the step of operating further comprising a step of operating each neural network device of the ensemble to provide an ensemble of outputs, the neural network device being trained to minimize the value representative of at least two said independent activation functions.

The system (200) for training an ensemble neural network device configured to execute the steps of: providing (205) a set of exemplar data, comprising at least one set of inputs (220) and at least one set of outputs (225) associated to the set of inputs, to a neural network device comprising an ensemble (230) of neural network devices, configured to provide independent predictions based upon the exemplar data, operating (210) the neural network device based upon the set of exemplar data, obtaining (215) the trained neural network device configured to provide an output, the neural network device further comprising at least two independent activation functions, whereof at least two of the independent activation functions are representative of the statistical distribution of the plurality of independent predictions, the neural network device being configured to provide at least one output (235, 236) for at least two said independent activation functions and the step of operating further comprising a step of operating each neural network device of the ensemble to provide an ensemble of outputs, the neural network device being trained to minimize the value representative of at least two said independent activation functions.

A computer-implemented method for predicting a conformation of a ligand docked into a protein is disclosed. According to some embodiments, the method may include determining one or more poses of the ligand in the protein, the poses being representative conformations of the ligand. The method may also include determining, using a neural network, energy scores of the poses. The method may further include determining a proper conformation for the docked ligand based on the energy scores.