US 11,694,769 B2Granted

Systems and methods for de novo peptide sequencing from data-independent acquisition using deep learning

AssigneeBioinformatics Solutions Inc.

InventorsBaozhen Shan, Ngoc Hieu Tran, Ming Li, Lei Xin, Rui Qiao, Xin Chen, Chuyi Liu

Life of this patent

Estimating the term…

Loading patent…

Abstract

The present systems and methods introduce deep learning to de novo peptide sequencing from tandem mass spectrometry data, and in particular mass spectrometry data obtained by data-independent acquisition. The systems and methods achieve improvements in sequencing accuracy over existing systems and methods and enables complete assembly of novel protein sequences without assisting databases. To sequence peptides from mass spectrometry data obtained by data-independent acquisition, precursor profiles representing intensities of one or more precursor ion signals associated with a precursor retention time and fragment ion spectra representing signals from fragment ions and fragment retention times are fed into a neural network.

Claims36

Dependent claims

A computer implemented system for de novo sequencing of a peptide from mass spectrometry data acquired by data-independent acquisition using neural networks, the computer implemented system comprising:

at least one memory and at least one processor configured to receive:
a first input representing at least one precursor profile, each precursor profile representing intensities of one or more precursor ion signals associated with a precursor retention time;
a second input representing a plurality of fragment ion spectra for each precursor profile, each fragment ion spectra representing:
signals from fragment ions generated from an associated precursor ion, and
a fragment retention time;
and
provide a plurality of layered computing nodes configured to form an artificial neural network for generating a probability measure for one or more candidates to a next amino acid in an amino acid sequence, the artificial neural network trained on mass spectrometry data containing retention time, a plurality of fragment ions peaks of sequences differing in length and differing by one or more amino acids;
wherein the plurality of layered nodes are configured to receive a mass spectrometry spectrum data base on the first and second inputs, the mass spectrometry spectrum data representing the at least one precursor profile and the fragment ion spectra, the plurality of layered nodes comprising at least one convolutional layer for filtering mass spectrometry spectrum data to detect fragment ion peaks;
and
wherein the processor is configured to:
receive an input prefix representing a determined amino acid sequence of the peptide,
provide the mass spectrometry spectrum data to the plurality of layered nodes,
identify a next amino acid based on a candidate next amino acid having a greatest probability measure based on the output of the artificial neural network and the mass spectrometry spectrum data of the peptide;
update the determined amino acid sequence with the next amino acid,
and generate an output signal representing a final determined sequence;
wherein the plurality of layered nodes receives a matrix data representing the mass spectrometry spectrum data, and output a probability measure vector;
and wherein the second input comprises a matrix data representing:
i. batch size,
ii. number of amino acids;
iii. ion types;
iv. number of fragment ion spectra associated with a precursor profile;
and
v. window size for filtering fragment ion peaks.

20 more elements

A method for de novo sequencing of a peptide from mass spectrometry data acquired by data-independent acquisition using neural networks, the method comprising:

receiving a first input representing at least one precursor profile, each precursor profile representing intensities of one or more precursor ion signals associated with a precursor retention time;
receiving a second input representing a plurality of fragment ion spectra for each precursor profile, each fragment ion spectra representing:
signals from fragment ions generated from an associated precursor ion, and
a fragment retention time;
filtering the mass spectrometry spectrum data to detect fragment ion peaks by at least one convolutional layer of a plurality of layered nodes configured to form an artificial neural network for generating a probability measure for one or more candidates to a next amino acid in an amino acid sequence;
receiving a probability measure for each candidate of a next amino acid;
obtaining an input prefix representing a determined amino acid sequence of the peptide;
providing a mass spectrometry spectrum data based on the first and second inputs to the plurality of layered nodes;
identifying a next amino acid based on a candidate next amino acid having a greatest probability measure based on the output of the artificial neural network and the mass spectrometry spectrum data of the peptide;
updating the determined amino acid sequence with the next amino acid;
and
generating an output signal representing a final determined sequence;
wherein the plurality of layered nodes receives a matrix data representing the mass spectrometry spectrum data, and output a probability measure vector;
and wherein the second input comprises a matrix data representing:
a) batch size,
b) number of amino acids;
c) ion types;
d) number of fragment ion spectra associated with a precursor profile;
and
e) window size for filtering fragment ion peaks.

17 more elements

A non-transitory computer readable media storing machine interpretable instructions, which when executed, cause a processor to perform steps of a method comprising:

receiving a first input representing at least one precursor profile, each precursor profile representing intensities of one or more precursor ion signals associated with a precursor retention time;
receiving a second input representing a plurality of fragment ion spectra for each precursor profile, each fragment ion spectra representing:
signals from fragment ions generated from an associated precursor ion, and
a fragment retention time;
filtering the mass spectrometry spectrum data to detect fragment ion peaks by at least one convolutional layer of a plurality of layered nodes configured to form an artificial neural network for generating a probability measure for one or more candidates to a next amino acid in an amino acid sequence;
receiving a probability measure for each candidate of a next amino acid;
obtaining an input prefix representing a determined amino acid sequence of the peptide;
providing a mass spectrometry spectrum data based on the first and second inputs to the plurality of layered nodes;
identifying a next amino acid based on a candidate next amino acid having a greatest probability measure based on the output of the artificial neural network and the mass spectrometry spectrum data of the peptide;
updating the determined amino acid sequence with the next amino acid;
and
generating an output signal representing a final determined sequence;
wherein the plurality of layered nodes receives a matrix data representing the mass spectrometry spectrum data, and output a probability measure vector;
and wherein the second input comprises a matrix data representing:
a) batch size,
b) number of amino acids;
c) ion types;
d) number of fragment ion spectra associated with a precursor profile;
and
e) window size for filtering fragment ion peaks.

17 more elements

Dependencies are inferred from the claim text.

Facts

Application: 16/226,575Filed 19 Dec 2018 · continuation in part
Priority: 17 Jul 2017
Published: US 2019/0018019 A116 May 2019
Granted: 4 Jul 2023
Term ends: Not estimatedThe source records don’t support an estimate.

Classification11 codes

G16B 40/10
ICT specially adapted for biostatistics; ICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge…: signal…
G01N 33/6818
General methods of protein analysis not limited to specific proteins or families of proteins: sequencing of polypeptides
G01N 33/6848
General methods of protein analysis not limited to specific proteins or families of proteins: methods of protein analysis involving mass spectrometry
G06N 3/02
Computing arrangements based on biological models: neural networks

Family

Application62/533,560Us provisional application · filed 17 Jul 2017
Application16/037,949Parent of this continuation-in-part · filed 17 Jul 2018
US 2019/0147983 A1Pre-grant publication · published 16 May 2019
US 11,573,239 B2Parent of this continuation-in-part
US 2019/0018019 A1Published application
US 11,694,769 B2This patent

From the US filing history. Foreign family members aren’t listed.