#METABOLOMICS WORKBENCH Safia_Firdous_20211026_002943 DATATRACK_ID:2907 STUDY_ID:ST002015 ANALYSIS_ID:AN003283 PROJECT_ID:PR001279 VERSION 1 CREATED_ON November 22, 2021, 10:12 pm #PROJECT PR:PROJECT_TITLE Dysregulated Alanine as a Potential Predictive Marker of Glioma—An Insight PR:PROJECT_TITLE from Untargeted HRMAS-NMR and Machine Learning Data PR:PROJECT_TYPE Untargeted HRMAS NMR, Glioma PR:PROJECT_SUMMARY Metabolic alterations play a crucial role in glioma development and progression PR:PROJECT_SUMMARY and can be detected even before the appearance of the fatal phenotype. We have PR:PROJECT_SUMMARY compared the circulating metabolic fingerprints of glioma patients versus PR:PROJECT_SUMMARY healthy controls, for the first time, in a quest to identify a panel of small, PR:PROJECT_SUMMARY dysregulated metabolites with potential to serve as a predictive and/or PR:PROJECT_SUMMARY diagnostic marker in the clinical settings. High-resolution magic angle spinning PR:PROJECT_SUMMARY nuclear magnetic resonance spectroscopy (HRMAS-NMR) was used for untargeted PR:PROJECT_SUMMARY metabolomics and data acquisition followed by a machine learning (ML) approach PR:PROJECT_SUMMARY for the analyses of large metabolic datasets. Cross-validation of ML predicted PR:PROJECT_SUMMARY NMR spectral features was done by statistical methods (Wilcoxon-test) using PR:PROJECT_SUMMARY JMP-pro16 software. Alanine was identified as the most critical metabolite with PR:PROJECT_SUMMARY potential to detect glioma with precision of 1.0, recall of 0.96, and F1 measure PR:PROJECT_SUMMARY of 0.98. The top 10 metabolites identified for glioma detection included PR:PROJECT_SUMMARY alanine, glutamine, valine, methionine, N-acetylaspartate (NAA), γ-aminobutyric PR:PROJECT_SUMMARY acid (GABA), serine, α-glucose, lactate, and arginine. We achieved 100% PR:PROJECT_SUMMARY accuracy for the detection of glioma using ML algorithms, extra tree classifier, PR:PROJECT_SUMMARY and random forest, and 98% accuracy with logistic regression. Classification of PR:PROJECT_SUMMARY glioma in low and high grades was done with 86% accuracy using logistic PR:PROJECT_SUMMARY regression model, and with 83% and 79% accuracy using extra tree classifier and PR:PROJECT_SUMMARY random forest, respectively. The predictive accuracy of our ML model is superior PR:PROJECT_SUMMARY to any of the previously reported algorithms, used in tissue- or liquid PR:PROJECT_SUMMARY biopsy-based metabolic studies. The identified top metabolites can be targeted PR:PROJECT_SUMMARY to develop early diagnostic methods as well as to plan personalized treatment PR:PROJECT_SUMMARY strategies. PR:INSTITUTE University of the Punjab PR:DEPARTMENT School of Biochemistry and Biotechnology PR:LABORATORY Biopharmaceuticals and Biomarkers Discovery Lab PR:LAST_NAME Firdous PR:FIRST_NAME Safia PR:ADDRESS Quaid e Azam Campus, University of the Punjab, Lahore. PR:EMAIL saima.ibb@pu.edu.pk PR:PHONE +924299231098 PR:FUNDING_SOURCE HEC-IRSIP, USA NIH grants: S10OD023406 and R21CA243255 PR:PUBLICATIONS Dysregulated Alanine as a Potential Predictive Marker of Glioma—An Insight PR:PUBLICATIONS from Untargeted HRMAS-NMR and Machine Learning Data PR:CONTRIBUTORS Safia Firdous, Rizwan Abid, Zubair Nawaz, Faisal Bukhari, Ammar Anwer, Leo L PR:CONTRIBUTORS Cheng, Saima Sadaf #STUDY ST:STUDY_TITLE Dysregulated Alanine as a Potential Predictive Marker of Glioma—An Insight ST:STUDY_TITLE from Untargeted HRMAS-NMR and Machine Learning Data ST:STUDY_TYPE Untargeted NMR ST:STUDY_SUMMARY Metabolic alterations play a crucial role in glioma development and progression ST:STUDY_SUMMARY and can be detected even before the appearance of the fatal phenotype. We have ST:STUDY_SUMMARY compared the circulating metabolic fingerprints of glioma patients versus ST:STUDY_SUMMARY healthy controls, for the first time, in a quest to identify a panel of small, ST:STUDY_SUMMARY dysregulated metabolites with potential to serve as a predictive and/or ST:STUDY_SUMMARY diagnostic marker in the clinical settings. High-resolution magic angle spinning ST:STUDY_SUMMARY nuclear magnetic resonance spectroscopy (HRMAS-NMR) was used for untargeted ST:STUDY_SUMMARY metabolomics and data acquisition followed by a machine learning (ML) approach ST:STUDY_SUMMARY for the analyses of large metabolic datasets. Crossvalidation of ML predicted ST:STUDY_SUMMARY NMR spectral features was done by statistical methods (Wilcoxon-test) using ST:STUDY_SUMMARY JMP-pro16 software. Alanine was identified as the most critical metabolite with ST:STUDY_SUMMARY potential to detect glioma with precision of 1.0, recall of 0.96, and F1 measure ST:STUDY_SUMMARY of 0.98. The top 10 metabolites identified for glioma detection included ST:STUDY_SUMMARY alanine, glutamine, valine, methionine, N-acetylaspartate (NAA), γ-aminobutyric ST:STUDY_SUMMARY acid (GABA), serine, α-glucose, lactate, and arginine. We achieved 100% ST:STUDY_SUMMARY accuracy for the detection of glioma using ML algorithms, extra tree classifier, ST:STUDY_SUMMARY and random forest, and 98% accuracy with logistic regression. Classification of ST:STUDY_SUMMARY glioma in low and high grades was done with 86% accuracy using logistic ST:STUDY_SUMMARY regression model, and with 83% and 79% accuracy using extra tree classifier and ST:STUDY_SUMMARY random forest, respectively. The predictive accuracy of our ML model is superior ST:STUDY_SUMMARY to any of the previously reported algorithms, used in tissue- or liquid ST:STUDY_SUMMARY biopsy-based metabolic studies. The identified top metabolites can be targeted ST:STUDY_SUMMARY to develop early diagnostic methods as well as to plan personalized treatment ST:STUDY_SUMMARY strategies. ST:INSTITUTE University of the Punjab ST:DEPARTMENT School of Biochemistry and Biotechnology ST:LABORATORY Biopharmaceuticals and Biomarkers Discovery Lab ST:LAST_NAME Firdous ST:FIRST_NAME Safia ST:ADDRESS Quaid e Azam Campus, University of the Punjab, Lahore. ST:EMAIL saima.ibb@pu.edu.pk ST:PHONE +924299231098 ST:NUM_GROUPS 2 ST:TOTAL_SUBJECTS 42 ST:NUM_MALES 25 ST:NUM_FEMALES 17 ST:PUBLICATIONS Dysregulated Alanine as a Potential Predictive Marker of Glioma—An Insight ST:PUBLICATIONS from Untargeted HRMAS-NMR and Machine Learning Data #SUBJECT SU:SUBJECT_TYPE Human SU:SUBJECT_SPECIES Homo sapiens SU:TAXONOMY_ID 9606 SU:AGE_OR_AGE_RANGE 15-60 Years SU:GENDER Male and female SU:HUMAN_RACE Asian SU:HUMAN_ETHNICITY Asian SU:HUMAN_LIFESTYLE_FACTORS N/A SU:HUMAN_MEDICATIONS N/A SU:HUMAN_PRESCRIPTION_OTC N/A SU:HUMAN_SMOKING_STATUS N/A SU:HUMAN_ALCOHOL_DRUG_USE N/A SU:HUMAN_NUTRITION N/A SU:HUMAN_INCLUSION_CRITERIA Low and High grade glioma patients confirmed by routine histopathology analysis SU:HUMAN_EXCLUSION_CRITERIA Diabetes mellitus, Hypertension, liver (hepatitis/liver cirrhosis), and SU:HUMAN_EXCLUSION_CRITERIA Cardiovascular disease #SUBJECT_SAMPLE_FACTORS: SUBJECT(optional)[tab]SAMPLE[tab]FACTORS(NAME:VALUE pairs separated by |)[tab]Raw file names and additional sample data SUBJECT_SAMPLE_FACTORS Glioma GBM1 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM2 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM3 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM4 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM5 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM6 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM7 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM8 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM9 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM10 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM11 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM12 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM13 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM14 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM15 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma GBM16 Class:HGG | Grade :IV SUBJECT_SAMPLE_FACTORS Glioma AA1 Class:HGG | Grade :III SUBJECT_SAMPLE_FACTORS Glioma DA1 Class:LGG | Grade :II SUBJECT_SAMPLE_FACTORS Glioma DA2 Class:LGG | Grade :II SUBJECT_SAMPLE_FACTORS Glioma DA3 Class:LGG | Grade :II SUBJECT_SAMPLE_FACTORS Glioma DA4 Class:LGG | Grade :II SUBJECT_SAMPLE_FACTORS Glioma DA5 Class:LGG | Grade :II SUBJECT_SAMPLE_FACTORS Glioma PA1 Class:LGG | Grade :I SUBJECT_SAMPLE_FACTORS Glioma PA2 Class:LGG | Grade :I SUBJECT_SAMPLE_FACTORS Glioma PA3 Class:LGG | Grade :I SUBJECT_SAMPLE_FACTORS Glioma PA4 Class:LGG | Grade :I SUBJECT_SAMPLE_FACTORS Control N1 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N2 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N3 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N4 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N5 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N6 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N7 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N8 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N9 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N10 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N11 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N12 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N13 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N14 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N15 Class:Control | Grade :- SUBJECT_SAMPLE_FACTORS Control N16 Class:Control | Grade :- #COLLECTION CO:COLLECTION_SUMMARY Peripheral blood (3 cc) from each patient (fasting state) was collected in CO:COLLECTION_SUMMARY Li-heparin tubes, centrifuged (300× g, 10 min) to prepare plasma within an hour CO:COLLECTION_SUMMARY of collection, and preserved in sterile tubes at −80 ◦C, as 200 µL CO:COLLECTION_SUMMARY aliquots, until further analyses. CO:SAMPLE_TYPE Blood (whole) CO:COLLECTION_LOCATION Punjab Institute of Neurosciences (PINS), Lahore, Pakistan. CO:COLLECTION_FREQUENCY Pre-operative CO:STORAGE_CONDITIONS -80℃ CO:COLLECTION_VIALS Li-Heparin #TREATMENT TR:TREATMENT_SUMMARY The enrolled patients underwent surgical resection of tumor after sample TR:TREATMENT_SUMMARY collection. #SAMPLEPREP SP:SAMPLEPREP_SUMMARY Sample was prepared by adding 10 µL plasma sample in a 4 mm zirconia rotor with SP:SAMPLEPREP_SUMMARY 12 µL Kel-F inserts; 2 µL D2O (Sigma Aldrich, St. Louis, MO, USA) with SP:SAMPLEPREP_SUMMARY reference trimethylsilylpropanoic acid (TSP) was added for field locking. SP:PROCESSING_STORAGE_CONDITIONS On ice #ANALYSIS AN:ANALYSIS_TYPE NMR AN:LABORATORY_NAME Martinos Center for Biomedical Imaging AN:OPERATOR_NAME Leo L Cheng AN:DETECTOR_TYPE Topspin AN:SOFTWARE_VERSION Bruker Biospin NMR System AN:ACQUISITION_DATE June 2018-January 2019 #NMR NM:INSTRUMENT_NAME Bruker Avence NM:INSTRUMENT_TYPE Other NM:NMR_EXPERIMENT_TYPE Other NM:NMR_COMMENTS Triple nucleus (1H,13C,31P) HRMAS probe NM:FIELD_FREQUENCY_LOCK D2O NM:SPECTROMETER_FREQUENCY 600MHz NM:NMR_PROBE Triple nucleus (1 H, 13 C, 31 P) HRMAS probe NM:NMR_SOLVENT D2O NM:NMR_TUBE_SIZE 4mm NM:SHIMMING_METHOD Autoshim NM:PULSE_SEQUENCE 90° Pulse Sequence NM:WATER_SUPPRESSION PLdB9 NM:PULSE_WIDTH 3 μs NM:POWER_LEVEL -14 dB NM:CHEMICAL_SHIFT_REF_CPD TSP NM:TEMPERATURE 4℃ NM:NUMBER_OF_SCANS 256 NM:DUMMY_SCANS 4 NM:RELAXATION_DELAY 5 s NM:SPECTRAL_WIDTH 12 ppm NM:NUM_DATA_POINTS_ACQUIRED 4096 NM:LINE_BROADENING 0.5Hz NM:CHEMICAL_SHIFT_REF_STD TSP at 0ppm, Lactate at 1.318ppm, Alanine at 1.468ppm NM:BINNED_INCREMENT 0.01 NM:BINNED_DATA_PROTOCOL_FILE N/A NM:NMR_RESULTS_FILE HRMAS_NMR_data_Glioma.txt UNITS:Peak Area #END