Color Genomics, Inc.

Color Health, Inc.

Ryan Barrett

Katsuya Noguchi

Krishna Pant

Nishant Bhat

Taylor Sittler

Othman Laraki

Wendy McKennon

Huy Hong

Elad Gil

Zhenghua Li

Kurt Smith

Zhengua Li

Jeroen Van Den Akker

Nish Bhat

Genetic-variant data is obtained that corresponds to one or more variants associated with a client. Each of the one or more variants corresponds to an instance of one or more bases positioned at one or more first positions in a first genetic sequence differ from corresponding one or more bases positioned in a reference genetic sequence. The first genetic sequence is a genetic sequence of the client. Sensor data is obtained that provides an indication of one or more characteristics of a current or past environment of the client. The genetic-variant data and the sensor data is processed to generate a disease-risk metric corresponding to a predicted risk of the client developing a particular disease. A communication is generated that is indicative of the disease-risk metric. The communication is transmitted to a remote device.

Communication generation using sparse indicators and sensor data

Techniques are provided for detecting copy number variations. Each sequence read of a set of sequence reads is aligned with a portion of a reference sequence. A coverage vector is generated that includes a plurality of elements, each element in the plurality of elements indicating a number of the set of sequence reads that were aligned to a particular position within the reference sequence. A normalization vector is accessed that was generated based on performance of a component analysis on a set of other coverage vectors corresponding to a set of other subjects. An adjusted coverage vector is generated using the coverage vector and normalization vector. One or more subject-specific normalization values are generated based on the coverage vector. One or more copy number variations are identified that corresponding to the sample using the adjusted coverage vector and the subject-specific normalization values.

Techniques, systems, and products for analyzing sparse indicators and sensor data and generating communications are disclosed. The sensors may be associated with or incorporated into devices that may automatically relay sensor data for use in analyses and communication generation.

Embodiments in the disclosure are directed to the use of distributed computing to align reads against multiple portions of a reference dataset. Aligned portions of the reference dataset that correspond with an above-threshold alignment score can be assessed for the presence of sparse indicators that can be categorized and used to influence a determination of a state transition likelihood. Various tasks associated with the processing of reads (e.g., alignment, sparse indicator detection, and/or determination of a state transition likelihood) may be able to take advantage of parallel processing and can be distributed among the machines while considering the resource utilization of those machines. Different load-balancing mechanisms can be employed in order to achieve even resource utilization across the machines, and in some cases may involve assessing various processing characteristics that reflect a predicted resource expenditure and/or time profile for each task to be processed by a machine.

Load balancing and conflict processing in workflow with task dependencies

A read is aligned to a reference data set. It is determined whether the read includes any identifier distinction, the determination being performed using the alignment. If so, positional data corresponding to the identifier distinction(s) are defined. Compressed read data is stored in association with a read identifier of the read. The compressed read data includes alignment information (e.g., a start and/or stop position of the alignment). When the read includes an identifier distinction, the compressed read data further includes the positional data and deviation data characterizing the distinction.

Lossless compression of client read data

Techniques, systems, and products for analyzing sparse indicators and generating communications based on bucketing of sparse indicators are disclosed.

Detecting and bucketing sparse indicators for communication generation

Techniques for using functional testing to detect run-time impacts of code modifications. A method includes accessing a workflow including a plurality of stages for processing reads. The stages are defined based on modifiable code and include a first stage for aligning reads with a corresponding portion of a reference data set and a second stage for collectively analyzing data corresponding to the aligned reads. The method includes identifying functional testing specifications to correspond with the workflow, including a definition of which stages are to be performed during functional testing, a reduced reference data set, and a set of reads. The method includes performing the functional testing using the reduced reference data set and the set of reads, detecting a result generated via the performance, and outputting the result.

Functional testing of code modifications for read processing systems

Techniques are disclosed for processing and aligning incomplete data. A stream of data is received from a data source including a plurality of reads. While receiving the stream of data and prior to having received all of the plurality of reads, a set of reads is extracted from the plurality of reads. Each of the set of reads is aligned to a corresponding portion of a reference data set. For each particular position of a plurality of particular positions of the reference data set, a subset of reads of the aligned set of reads is identified. A value of a client data set is generated based on the subset of reads. A variable is generated based on the client data set. Data is routed when a condition, based on the variable, is satisfied.

Real-time alignment and processing of incomplete stream of data

Methods and systems disclosed herein relate generally to data processing by applying machine learning techniques to iteration data to identify anomaly subsets of iteration data. More specifically, iteration data for individual iterations of a workflow involving a set of tasks may contain a client data set, client-associated sparse indicators and their classifications, and a set of processing times for the set of tasks performed in that iteration of the workflow. These individual iterations of the workflow may also be associated with particular data sources. Using the iteration data, anomaly subsets within the iteration data can be identified, such as data items resulting from systematic error associated with particular data sources, sets of sparse indicators to be validated or double-checked, or tasks that are associated with long processing times. The anomaly subsets can be provided in a generated communication or report in order to optimize future iterations of the workflow.

Techniques for processing queries relating to task-completion times or cross-data-structure interactions

Methods and systems disclosed herein relate generally to processing data requests from external assessment systems. More specifically, an interface is availed to external assessment systems that accepts an identification of one or more genes. Upon receiving a request identifying one or more genes, a type of access authorized for the requesting external assessment system is assessed. When it is determined that the type of data access indicates that the external assessment system is authorized to access data for the one or more genes, a data repository is queried to identify client data that corresponds to the one or more genes and that indicates or can be used to detect a presence of client-associated variants. A response data set that includes at least some of the client data is transmitted to the external assessment system.

Systems and methods for processing requests for genetic data based on client permission data

Techniques for accurately identifying duplications and deletions using depth vectors. A depth vector is generated for each of multiple clients based on a set of reads that is received and aligned to a reference data set. A transformation processing of the depth vectors is performed to produce multiple components. Each of the components is assigned an order based on the extent to which it accounts for cross-client differences in the depth vectors. Each of the components includes an intensity, multiple values, and multiple client weights. A subset of the components is identified based on the order. A sparse indicator and positional data for the sparse indicator can be determined from the components in the subset, and one or more clients can be identified as being associated with the components.

Growing community of inventors

San Francisco, CA, United States of America

Ryan Barrett