Sample preparation for glycoproteomic analysis that includes diagnosis of disease

A method for proteolytically digesting a biological sample comprising a glycoprotein to produce a proteolytic glycopeptide, the method comprising:

• subjecting the biological sample to a thermal denaturation technique to produce a denatured sample,
• wherein the thermal denaturation technique comprises subjecting the biological sample to a thermal cycle comprising a thermal treatment of about 60° C. to about 100° C. with a hold time of at least about 1 minute,
• wherein a lid temperature during the thermal cycle is at least about 2° C. higher than a temperature of a block temperature during the thermal cycle;
• subjecting the denatured sample to a reduction technique to produce a reduced sample, wherein the reduction technique comprises adding an amount of a reducing agent to the denatured sample and incubating for a reducing incubation time;
• subjecting the reduced sample to an alkylation technique to produce an alkylated sample, wherein the alkylation technique comprises adding an amount of an alkylating agent to the reduced sample and incubating substantially in a dark or in a low light condition for an alkylation incubation time, and
• wherein the alkylated technique comprises quenching the alkylating agent following the alkylation incubation time;
• and
• subjecting the alkylated sample to a proteolytic digestion technique to produce a proteolytically digested sample comprising the proteolytic glycopeptide,
• wherein the proteolytic digestion technique comprises adding an amount of one or more proteolytic enzymes and incubating for a digestion incubation time, and
• wherein the proteolytic digestion technique comprises quenching the one or more proteolytic enzymes following the digestion incubation time.

A method for processing a proteolytically digested sample to produce a processed sample suitable for use in a liquid chromatography-mass spectrometry (LC-MS) analysis, wherein the proteolytically digested sample comprises a plurality of proteolytic polypeptides comprising at least one proteolytic glycopeptide,

• the method comprising:
• performing one or more of the following:
• (a) subjecting the proteolytically digested sample to a solid phase extraction column comprising a reversed-phase medium according to one or more conditions to associate at least a portion of the plurality of proteolytic polypeptides with the reversed-phase medium, the one or more conditions comprising:
• (i) a polypeptide loading amount of about 50% or less of a binding capacity of the reversed-phase medium,
• wherein the binding capacity of the reversed-phase medium is based on an insulin load having 10% or less breakthrough;
• or
• (ii) a polypeptide loading concentration of about 0.6 μg/μL or less;
• or
• (b) subjecting the reversed-phase medium comprising the associated proteolytic polypeptides to a wash buffer at a wash flow rate of about 0.1 column volumes/minute to about 2 column volumes/minute;
• and
• subjecting the reversed-phase medium comprising the associated proteolytic polypeptides to an elution buffer to produce the processed sample.

A method for performing a liquid chromatography-mass spectrometry (LC-MS) analysis of a proteolytic glycopeptide derived from a blood sample deposited on a delimited zone of an absorbent or bibulous member

• wherein the blood sample comprises a plurality of polypeptides comprising at least one glycoprotein,
• the method comprising:
• extracting at least a portion of the plurality of polypeptides and one or more extraction internal standards from the absorbent or bibulous member to obtain an extracted sample,
• wherein the absorbent or bibulous member comprises the one or more extraction internal standards prior to deposition of the blood sample within the delimited zone, and
• wherein at least one of the one or more extraction internal standards comprises a polypeptide standard;
• subjecting the extracted sample or a derivative thereof to a proteolytic digestion technique to produce a proteolytically digested sample comprising the proteolytic glycopeptide;
• introducing at least a portion of the proteolytically digested sample to a liquid chromatography (LC) system of a LC-MS system;
• and
• performing the LC-MS analysis on at least the proteolytic glycopeptide and the one or more extraction internal standards.

A method of classifying a biological sample obtained from a subject with respect to a plurality of states associated with a pelvic cancer, the method comprising receiving peptide structure data corresponding to a set of glycoproteins in the biological sample;

• inputting quantification data identified from the peptide structure data for a set of peptide structures into a machine-learning model trained to identify a disease indicator based on the quantification data, wherein the set of peptide structures comprises at least one peptide structure identified from a plurality of peptide structures in Table 9;
• identifying, by the machine-learning model, the disease indicator;
• and
• classifying the biological sample with respect to a plurality of states associated with pelvic cancer based upon the identified disease indicator.

A method of detecting the presence of one of a plurality of states associated with a pelvic cancer in a subject, the method comprising

• receiving peptide structure data corresponding to a set of glycoproteins in a biological sample obtained from a subject, wherein the peptide structure data comprises at least one peptide structure from Table 9;
• inputting quantification data identified from the peptide structure data for a set of peptide structures into a machine-learning model trained to identify a disease indicator based on the quantification data;
• and
• detecting the presence of a corresponding state of the plurality of states associated with the pelvic cancer in response to a determination that the identified disease indicator falls within a selected range associated with the corresponding state.

A method of treating a pelvic tumor in a subject comprising

• receiving peptide structure data corresponding to a set of glycoproteins in a biological sample obtained from a subject, wherein the peptide structure data comprises at least one peptide structure from Table 9;
• inputting quantification data for the at least one peptide structure into a machine-learning model trained to generate a risk score based on the quantification data;
• outputting, by the machine-learning model, the quantification data using the machine learning model to generate a risk score,
• administering an effective amount of an agent to treat the pelvic cancer based upon the risk score.

A method of determining a diagnosis for a pelvic tumor in a subject comprising receiving peptide structure data corresponding to a set of glycoproteins in a biological sample;

• inputting quantification data identified from the peptide structure data for a set of peptide structures into a machine-learning model trained to identify a disease indicator based on the quantification data, wherein the set of peptide structure data comprises at least one peptide structure identified from a plurality of peptide structures in Table 9;
• identifying, by the machine-learning model, the disease indicator;
• and
• determining a diagnosis for the pelvic tumor based upon the identified disease indicator.

A method of treating a pelvic tumor in a subject comprising

• receiving peptide structure data corresponding to a set of glycoproteins in a biological sample;
• inputting quantification data identified from the peptide structure data for a set of peptide structures into a machine-learning model trained to identify a disease indicator based on the quantification data, wherein the peptide structure data comprises at least one peptide structure identified from a plurality of peptide structures in Table 9;
• identifying, by the machine-learning model, the disease indicator;
• determining a risk score the identified disease indicator;
• and
• administering an effective amount of an agent to treat the pelvic tumor based upon the risk score.

A method of diagnosing an individual with a pelvic tumor comprising

• detecting the presence or amount of at least one peptide structure from Table 9;
• inputting a quantification of the detected at least one peptide structure into a machine-learning model trained to generate a class label,
• determining if the class label is above or below a threshold for a classification;
• identifying a diagnostic classification for the individual based on whether the class label is above or below a threshold for the classification;
• and
• diagnosing the individual as having a benign or malignant pelvic tumor on the diagnostic classification.

A method for performing a liquid chromatography-mass spectrometry (LC-MS) analysis of a proteolytic glycopeptide derived from a blood sample from an individual deposited on a delimited zone of a blood spot card, the method comprising

• obtaining a blood spot card comprising a blood sample from the individual deposited thereon, wherein the blood spot card comprises one or more extraction internal standards deposited and dried prior to deposition of the blood sample on the blood spot card, and
• wherein the blood spot card comprising the blood sample contains at least a portion of the blood sample and the one or more extraction internal standards in an overlapping area of the blood spot card;
• extracting at least a portion of the plurality of polypeptides and the one or more extraction internal standards from the blood spot card to obtain an extracted sample;
• subjecting the extracted sample or a derivative thereof to a proteolytic digestion technique to produce a proteolytically digested sample comprising the proteolytic glycopeptide;
• introducing at least a portion of the proteolytically digested sample to a liquid chromatography (LC) system of a LC-MS system;
• and
• performing an LC-MS analysis to quantify one or more biomarkers of ovarian cancer and the one or more extraction internal standards,
• wherein the one or more biomarkers comprise a polypeptide comprising a sequence of any of SEQ ID NOs:
• 35-51, and
• wherein at least one of the one or more biomarkers is a glycopeptide.

A method of training a model to diagnose a subject with one of a plurality of states associated with a pelvic tumor, the method comprising

• receiving quantification data for a panel of peptide structures for a plurality of subjects
• diagnosed with the plurality of states associated with a pelvic tumor
• wherein the panel of peptide structures comprises at least one peptide structure set forth in Table 9;
• and
• training a machine-learning model to determine a state of the plurality of states a biological sample from the subject based on the quantification data.

A method for processing a proteolytic digest sample for use in a liquid chromatography-mass spectrometry (LC-MS) analysis,

• wherein the proteolytic digest sample comprises a plurality of proteolytically digested peptides comprising at least one proteolytically digested glycopeptide,
• the method comprising:
• (A) loading a hydrophilic interaction liquid chromatography (HILIC) load derived from the proteolytic digest sample to a solid phase extraction column comprising a HILIC medium according to one or more conditions to associate the at least one proteolytically digested glycopeptide with the HILIC medium, the one or more conditions comprising:
• (1) the loading of the HILIC load to the solid phase extraction column is initiated when the HILIC medium is in a dry state;
• (2) the HILIC load loaded to the solid phase extraction column has an amount of the plurality of proteolytically digested peptides characterized by one or both of:
• (a) a ratio of a weight of the plurality of proteolytically digested peptides over a weight of the HILIC medium in the dry state of at least about 0.06;
• and/or
• (b) a ratio of the weight of the plurality of proteolytically digested peptides relative to a bed volume of the HILIC medium in the dry state of at least about 40 μg/μl;
• or
• (3) the HILIC load loaded to the solid phase extraction column has a concentration of an organic solvent of at least about 70% (v/v);
• and
• (B) subjecting the HILIC medium to an elution liquid to obtain a HILIC eluate comprising the at least one proteolytically digested glycopeptide.

A method of processing a blood-derived sample obtained from an individual for a glycoproteomic mass spectrometry (MS) technique, the method comprising:

• (a) admixing the blood-derived sample with one or more defibrination factors to promote formation of a fibrin clot, the one or more defibrination factors comprises one or more members selected from the group consisting of:
• a clotting co-factor;
• a clotting enzyme;
• and
• a clotting activator and/or an exogenous surface aggregation agent;
• (b) separating the formed fibrin clot from the admixed blood-derived sample to obtain a fibrinogen-depleted sample;
• and
• (c) subjecting the fibrinogen-depleted sample to one or more MS preparation techniques to produce a test sample for the glycoproteomic mass spectrometry technique.

A method of preparing a plasma sample obtained from an individual for a glycoproteomic mass spectrometry technique, the method comprising:

• (a) admixing the plasma sample with defibrination factors to promote formation of a fibrin clot, the defibrination factors comprising:
• a clotting co-factor;
• a clotting enzyme;
• and
• a clotting activator and/or an exogenous surface aggregation agent;
• (b) separating the formed fibrin clot from the admixed plasma sample to obtain a fibrinogen-depleted sample;
• and
• (c) subjecting the fibrinogen-depleted sample to one or more MS preparation techniques to produce a test sample for the glycoproteomic mass spectrometry technique.

A defibrination composition comprising:

• a clotting co-factor;
• a clotting enzyme;
• and
• a clotting activator and/or an exogenous surface aggregation agent.

A method of monitoring a subject for a melanoma, the method comprising:

• receiving first monomer weight score data for a first biological sample obtained from a subject at a first timepoint;
• analyzing the first monomer weight score data using at least one supervised machine learning model to generate a first disease indicator based on at least one site monomer selected from a group of site monomers identified in Table 16, wherein the group of site monomers in Table 16 comprises a group of site monomers having monomer weight scores associated with melanoma;
• receiving second monomer weight score data of a second biological sample obtained from the subject at a second timepoint;
• analyzing the second monomer weight score data using the at least one supervised machine learning model to generate a second disease indicator based on the at least one site monomer selected from the group of site monomers identified in Table 16;
• and
• generating a diagnosis output based on the first disease indicator and the second disease indicator.

A method of treating melanoma in a subject, the method comprising:

• determining a monomer weight score for at least one site monomer identified in Table 16 in a biological sample from the subject using a multiple reaction monitoring mass spectrometry (MRM-MS) system;
• analyzing the monomer weight score using at least one machine learning model to generate a disease indicator;
• generating a diagnosis output based on the disease indicator that classifies the biological sample as evidencing that the patient has melanoma;
• and
• administering to the subject a therapeutically effective amount of a melanoma therapy.

A method of treating melanoma in a subject, the method comprising:

• determining a monomer weight score for at least one site monomer identified in Table 16 in a biological sample from the subject using a multiple reaction monitoring mass spectrometry (MRM-MS) system;
• analyzing the monomer weight score using at least one machine learning model to generate a disease indicator;
• generating a diagnosis output based on the disease indicator that classifies the biological sample as evidencing that the melanoma is sensitive to immunotherapy;
• and
• administering to the subject a therapeutically effective amount of immunotherapy.

A method of predicting a risk for melanoma in a subject, the method comprising:

• determining a monomer weight score for at least one site monomer identified in Table 16 in a biological sample from the subject using a multiple reaction monitoring mass spectrometry (MRM-MS) system;
• analyzing the monomer weight score using at least one machine learning model to generate a disease indicator;
• and
• generating a diagnosis output based on the disease indicator that classifies the biological sample as evidencing that the patient has a risk for melanoma.

A method of predicting immunotherapy sensitivity, the method comprising:

• determining a monomer weight score for at least one site monomer identified in Table 16 in a biological sample from the subject using a multiple reaction monitoring mass spectrometry (MRM-MS) system;
• analyzing the monomer weight score using at least one machine learning model to generate a disease indicator;
• and
• generating a diagnosis output based on the disease indicator that classifies the biological sample as evidencing that the patient has a risk for melanoma.

A method of classifying a biological sample with respect to a responsiveness to immune checkpoint inhibitor therapy, the method comprising:

• A) analyzing one or more monomer weight scores of a set of peptide structures from a biological sample from the subject using a machine learning model to generate a disease indicator;
• B) generating a diagnosis output based on the disease indicator that classifies the biological sample as evidencing the responsiveness to immune checkpoint inhibitory therapy, wherein the one or more monomer weight scores correspond to at least one site monomer identified in Table 29;
• and
• C) administering to the subject a therapeutically effective amount of immunotherapy.

A method for predicting retention times of peptides, by a computing system comprising one or more processors:

• accessing a feature set corresponding to a peptide, wherein the feature set represents peptide sequence data of the peptide and corresponding physicochemical features;
• sending the feature set as an input into a neural network, the neural network comprising:
• (1) a plurality of 1DCNN layers, (2) one or more BiLSTM layers, and (3) a multi-head attention layer;
• and
• obtaining, as an output from the neural network, a predicted retention time for the peptide corresponding to an estimated retention time for the peptide in a liquid chromatography mass spectrometry (LC-MS) run.

A method of training a neural network for predicting retention times of peptides, by a computing system comprising one or more processors:

• accessing a plurality of feature sets corresponding to a plurality of peptides, wherein the feature set represents peptide sequence data of the peptide and corresponding physicochemical features;
• creating a training set comprising a subset of feature sets from the plurality of feature sets;
• and
• training a neural network using the training set, the neural network comprising:
• (1) a plurality of 1DCNN layers, (2) one or more BiLSTM layers, and (3) a multi-head attention layer.

A system for predicting retention times of peptides, the system comprising:

• one or more processors;
• and
• a non-transitory memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:
• access a feature set corresponding to a peptide, wherein the feature set represents peptide sequence data of the peptide and corresponding physicochemical features;
• send the feature set as an input into a neural network, the neural network comprising:
• (1) a plurality of 1DCNN layers, (2) one or more BiLSTM layers, and (3) a multi-head attention layer. obtain, as an output from the neural network, a predicted retention time for the peptide corresponding to an estimated retention time for the peptide in a liquid chromatography mass spectrometry (LC-MS) run.

A system for training a neural network for predicting retention times of peptides, the system comprising:

• one or more processors;
• and
• a non-transitory memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:
• access a plurality of feature sets corresponding to a plurality of peptides, wherein the feature set represents peptide sequence data of the peptide and corresponding physicochemical features;
• create a training set comprising a subset of feature sets from the plurality of feature sets;
• and
• train a neural network using the training set, the neural network comprising:
• (1) a plurality of 1DCNN layers, (2) one or more BiLSTM layers, and (3) a multi-head attention layer.