US Patent 7,103,222. Pattern discovery in multi-dimensional time series using multi-resolution matching. Inventor: Kadir Peker. Assignee: Mitsubishi Electric Research Laboratories.
Protects a method for discovering patterns in multidimensional data. The method generates a time-series of the multidimensional data and constructs a point cross-distance matrix by self-correlating the time-series. All minimum cost paths in the point cross-distance matrix are located at multiple time resolutions and the minimum cost paths are then “related to temporal sub-sequences in the multi-dimensional data to discover high-level patterns in the unknown multi-dimensional data,” according to the patent abstract.
US Patent 7,099,502. System and method for automatically processing microarrays. Inventors: Soheil Shams, James Darrell Park, Yi-Xiong Zhou. Assignee: BioDiscovery.
Protects a digital image processing-based system and method for quantitatively processing nucleic acid species expressed in a microarray. The system includes a scanner with a digital scanning sensor that scans the microarray and transmits from an output a digital image of the microarray, and a processor that receives the digital image from the scanner and then it, identifying each of the microarray's sub-grids. The processor then detects in each of the sub-grids a center-representing pixel of a signal of a chemical material and an approximate radius of the signal. The processor segments the signal and calculates a characterizing measure for the segmented signal.
US Patent 7,096,206. Heuristic method of classification. Inventor: Ben Hitt. Assignee: Correlogic Systems.
Covers heuristic algorithms for the classification of objects that are useful in the areas of “classifying texts and medical samples,” among other applications, according to the patent abstract. The method comprises a genetic algorithm that is used to abstract a data stream associated with each object and a pattern-recognition algorithm that is used to classify the objects and measure the fitness of the chromosomes of the genetic algorithm. The learning algorithm is applied to a training data set and generates a classifying algorithm, which is used to classify or categorize unknown objects.
US Patent 7,096,167. System and method for molecular dynamic simulation. Inventors: Ruhong Zhou, Bruce Berne, Edward Harder. Assignee: IBM.
Protects a system and method for molecular dynamic simulation. The method includes a database for storing data pertaining to at least one biomolecular system; a memory device for storing instructions for performing at least one algorithm having an electrostatic interaction calculating function and a multiple time step function; and a processor for processing the data. The system and method “significantly speed up the molecular dynamics simulation of biomolecular systems in which there are long-range and short-range electrostatic interactions and in which there are fast and slow motions, and make practicable the simulation of large protein solutions and thus, can be used to simulate protein folding and the binding of substrates to protein molecules,” according to the patent abstract.
US Patent 7,096,162. Computer-implemented method of merging libraries of molecules using validated molecular structural descriptors and neighborhood distances to maximize diversity and minimize redundancy. Inventors: Richard Cramer, David Patterson.
Protects a validation method that selects a subset of a combinatorial accessible chemical universe “such that the molecules of the subset are representative of all the diversity present in the universe and yet do not contain multiple members which represent the same diversity (oversample).” The method can be used to combine any number of combinatorial screening libraries, according to the patent abstract.
US Patent 7,089,168. Language for networks. Inventor: Ron Maimon. Assignee: Gene Network Sciences.
Protects a formal language called the Cell Language for describing the function of biochemical networks. The language includes a complex recursive algorithm for parsing. Cell Language is described both informally, so that it may be written, and formally, so that it may be parsed. “The Cell Language makes it possible to model all the interactions in a cell in a single diagram, with only a few representations of each molecule,” according to the patent abstract.
US Patent 7,085,652. Methods for searching polynucleotide probe targets in databases. Inventors: Charles Scaf, Nicolas Peyret, Francisco De La Vega, Ryan Koehler, Eugene Spier. Assignee: Applera.
Protects methods for determining the binding affinity of a probe to a target or targets in a polynucleotide composite. The methods determine relative binding sites based on thermodynamic principles, using a thermodynamic alphabet and a thermodynamic scoring matrix, with appropriate computer software, such as BLASTP. The invention also provides methods for designing polynucleotide probes to be used in hybridization assays, and minimizing the occurrence of cross-hybridization.
US Patent 7,085,651. Method and device for assembling nucleic acid base sequences. Inventors: Tomohiro Yasuda, Tetsuo Nishikawa. Assignee: Hitachi.
Covers a method for clustering and assembling nucleic acid base sequences at a high speed. The method moves an alignment window of a fixed base length along a first nucleic acid base sequence to define a first fixed-length sequence and simultaneously searches for a second nucleic acid base sequence having a terminal region matching exactly with the window of fixed base length. The method then determines whether the second nucleic acid base sequence and the first nucleic acid base sequence can be assembled by comparing a sequence adjacent to the aligned window of the first nucleic acid base sequence with a sequence adjacent to the aligned window of the second nucleic acid base sequence. The first and second nucleic acid base sequences are assembled if the sequences adjacent to the aligned windows are similar.