Back to results
US 2024/0257916 A1Published application
Automated detection of boundaries in mass spectrometry data
AssigneeVenn Biosciences Corporation
InventorsDaniel SERIE, Zhenqin Wu
Life of this patent
Estimating the term…
Abstract
A system and method for automated detection of the presence or absence of a quantity based on intensities expressed in terms of, or derived from frequency or time dependent data. According to one example intensities from mass spectrometry are identified using a non-linear mathematical model, such as an artificial neural network trained to find start and stop peaks of an intensity, from which an abundance may be determined.
Claims17
1-10 (canceled)
A method, comprising:
- featurizing data from raw MS data, wherein the raw MS data comprises a plurality of analog samples, each representing mass/charge intensity, the featurizing comprising,
- for each of the analog samples, the steps of
- centering the sample within a retention time window;
- discretizing the sample into a sequence of points representing intensity values;
- standardizing the intensity values;
- and
- assigning labels to points among the sequence of points corresponding to peak start and peak stop times to produce a labeled sequence of points;
- wherein the labeled sequence of points are configured for training a machine learning model to predict an abundance in unseen MS data.
5 more elements
16-18 (canceled)
A method comprising:
- using a plurality of labeled sequential data;
- and
- training at least one machine learning model using the labeled sequential data, wherein the trained machine learning model is adapted for identifying a time period or frequency range where peak start and peak stop points of an intensity occur in an unseen sample.
A method comprising:
- providing at least one trained neural network model that produces an output as a function of input parameters, wherein values for the input parameters are a temporal or spectral sequence of intensity values representing a quantity;
- and
- providing values for each of the input parameters, the values representing an unseen sequence of intensities for the quantity;
- wherein the output comprises one or more peak start and peak stop times or frequencies, which are used to quantify the quantity.
1 more element
Dependencies are inferred from the claim text.