#METABOLOMICS WORKBENCH yuewu_20230524_075258 DATATRACK_ID:4040 STUDY_ID:ST002718 ANALYSIS_ID:AN004407 PROJECT_ID:PR001685
VERSION             	1
CREATED_ON             	May 25, 2023, 8:04 pm
#PROJECT
PR:PROJECT_TITLE                 	SAND: automated time-domain modeling of NMR spectra applied to metabolic
PR:PROJECT_TITLE                 	quantification
PR:PROJECT_TYPE                  	NMR quantification of spike-in samples
PR:PROJECT_SUMMARY               	New developments in untargeted nuclear magnetic resonance (NMR) metabolomics
PR:PROJECT_SUMMARY               	enable the profiling of hundreds to thousands of biological samples in
PR:PROJECT_SUMMARY               	biomedical studies, with great potential in drug discovery and diagnostics. The
PR:PROJECT_SUMMARY               	exploitation of this rich information requires detailed quantification of
PR:PROJECT_SUMMARY               	spectral features. However, the development of a consistent and automatic
PR:PROJECT_SUMMARY               	workflow for NMR feature quantification has been a long-standing challenge
PR:PROJECT_SUMMARY               	because of the difficulties of extensive spectral overlap. To address this
PR:PROJECT_SUMMARY               	challenge, we introduce the software SAND (Spectral Automated NMR
PR:PROJECT_SUMMARY               	Deconvolution), for automated feature quantification in the time domain. SAND
PR:PROJECT_SUMMARY               	follows upon the previous success of time-domain modeling and provides automated
PR:PROJECT_SUMMARY               	quantification of entire spectra without the need for manual interaction. SAND
PR:PROJECT_SUMMARY               	employs subsampling, global optimization, and statistic model selection, which
PR:PROJECT_SUMMARY               	are readily expandable to higher dimensional NMR and non-uniform sampling
PR:PROJECT_SUMMARY               	applications. Here, we demonstrate the accuracy of the SAND approach (a
PR:PROJECT_SUMMARY               	correlation around 0.9) using highly overlapped simulated datasets, a
PR:PROJECT_SUMMARY               	two-compound mixture, and a urine spectral series spiked with differing amounts
PR:PROJECT_SUMMARY               	of a four-compound mixture. We further demonstrate automated annotation using
PR:PROJECT_SUMMARY               	correlation networks derived from SAND deconvoluted peaks, and on average 74% of
PR:PROJECT_SUMMARY               	peaks for each compound can be recovered in a single correlation network
PR:PROJECT_SUMMARY               	cluster. SAND is currently integrated with NMRbox and the Network for Advanced
PR:PROJECT_SUMMARY               	NMR (NAN).
PR:INSTITUTE                     	University of Georgia
PR:DEPARTMENT                    	Genetics; Biochemistry and Molecular Biology; Institute of Bioinformatics;
PR:DEPARTMENT                    	College of Engineering; Complex Carbohydrate Research Center
PR:LABORATORY                    	Arthur S. Edison and Frank Delaglio
PR:LAST_NAME                     	Wu
PR:FIRST_NAME                    	Yue
PR:ADDRESS                       	3165 Porter Drive, Palo Alto, CA, 94304
PR:EMAIL                         	yuewu.mike@gmail.com
PR:PHONE                         	7062546619
PR:FUNDING_SOURCE                	NSF 1946970, NIH P41GM111135 (NIGMS)
PR:PUBLICATIONS                  	to be submitted soon
PR:CONTRIBUTORS                  	Yue Wu, Omid Sanati, Mario Uchimiya, Krish Krishnamurthy, Arthur S. Edison,
PR:CONTRIBUTORS                  	Frank Delaglio
#STUDY
ST:STUDY_TITLE                   	SAND: automated time-domain modeling of NMR spectra applied to metabolic
ST:STUDY_TITLE                   	quantification
ST:STUDY_TYPE                    	spike-in urine sample
ST:STUDY_SUMMARY                 	New developments in untargeted nuclear magnetic resonance (NMR) metabolomics
ST:STUDY_SUMMARY                 	enable the profiling of hundreds to thousands of biological samples in
ST:STUDY_SUMMARY                 	biomedical studies, with great potential in drug discovery and diagnostics. The
ST:STUDY_SUMMARY                 	exploitation of this rich information requires detailed quantification of
ST:STUDY_SUMMARY                 	spectral features. However, the development of a consistent and automatic
ST:STUDY_SUMMARY                 	workflow for NMR feature quantification has been a long-standing challenge
ST:STUDY_SUMMARY                 	because of the difficulties of extensive spectral overlap. To address this
ST:STUDY_SUMMARY                 	challenge, we introduce the software SAND (Spectral Automated NMR
ST:STUDY_SUMMARY                 	Deconvolution), for automated feature quantification in the time domain. SAND
ST:STUDY_SUMMARY                 	follows upon the previous success of time-domain modeling and provides automated
ST:STUDY_SUMMARY                 	quantification of entire spectra without the need for manual interaction. SAND
ST:STUDY_SUMMARY                 	employs subsampling, global optimization, and statistic model selection, which
ST:STUDY_SUMMARY                 	are readily expandable to higher dimensional NMR and non-uniform sampling
ST:STUDY_SUMMARY                 	applications. Here, we demonstrate the accuracy of the SAND approach (a
ST:STUDY_SUMMARY                 	correlation around 0.9) using highly overlapped simulated datasets, a
ST:STUDY_SUMMARY                 	two-compound mixture, and a urine spectral series spiked with differing amounts
ST:STUDY_SUMMARY                 	of a four-compound mixture. We further demonstrate automated annotation using
ST:STUDY_SUMMARY                 	correlation networks derived from SAND deconvoluted peaks, and on average 74% of
ST:STUDY_SUMMARY                 	peaks for each compound can be recovered in a single correlation network
ST:STUDY_SUMMARY                 	cluster. SAND is currently integrated with NMRbox and the Network for Advanced
ST:STUDY_SUMMARY                 	NMR (NAN).
ST:INSTITUTE                     	University of Georgia
ST:DEPARTMENT                    	Genetics; Biochemistry and Molecular Biology; Institute of Bioinformatics;
ST:DEPARTMENT                    	College of Engineering; Complex Carbohydrate Research Center
ST:LABORATORY                    	Arthur S. Edison and Frank Delaglio
ST:LAST_NAME                     	Wu
ST:FIRST_NAME                    	Yue
ST:ADDRESS                       	3165 Porter Drive
ST:EMAIL                         	yuewu.mike@gmail.com
ST:PHONE                         	7062546619
ST:NUM_GROUPS                    	n/a
ST:TOTAL_SUBJECTS                	n/a
ST:NUM_MALES                     	n/a
ST:NUM_FEMALES                   	n/a
ST:PUBLICATIONS                  	to be submitted
#SUBJECT
SU:SUBJECT_TYPE                  	Human
SU:SUBJECT_SPECIES               	Homo sapiens
SU:TAXONOMY_ID                   	9606
#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           	-	1	Run:Baseline | Urine (µL):540	Buffer (µL)=60; Spike-in (µL)=0; Note=-; RAW_FILE_NAME=1
SUBJECT_SAMPLE_FACTORS           	-	2	Run:Spike-in_1 | Urine (µL):540	Buffer (µL)=60; Spike-in (µL)=20; Note=20 µL of the spike-in solution was added to the tube after Run 1; RAW_FILE_NAME=2
SUBJECT_SAMPLE_FACTORS           	-	3	Run:Spike-in_2 | Urine (µL):540	Buffer (µL)=60; Spike-in (µL)=40; Note=20 µL of the spike-in solution was added to the tube after Run 2; RAW_FILE_NAME=3
SUBJECT_SAMPLE_FACTORS           	-	4	Run:Spike-in_3 | Urine (µL):540	Buffer (µL)=60; Spike-in (µL)=60; Note=20 µL of the spike-in solution was added to the tube after Run 3; RAW_FILE_NAME=4
SUBJECT_SAMPLE_FACTORS           	-	5	Run:Spike-in_4 | Urine (µL):540	Buffer (µL)=60; Spike-in (µL)=80; Note=20 µL of the spike-in solution was added to the tube after Run 4; RAW_FILE_NAME=5
SUBJECT_SAMPLE_FACTORS           	-	6	Run:Reference | Urine (µL):0	Buffer (µL)=60; Spike-in (µL)=20; Note=D2O_540µL+buffer_60µL+spike20uL; RAW_FILE_NAME=6
SUBJECT_SAMPLE_FACTORS           	-	7	Run:Blank | Urine (µL):0	Buffer (µL)=60; Spike-in (µL)=0; Note=D2O_540µL+buffer_60µL; RAW_FILE_NAME=7
#COLLECTION
CO:COLLECTION_SUMMARY            	A spike-in experiment was conducted by adding a mixture of compounds to a urine
CO:COLLECTION_SUMMARY            	sample.
CO:COLLECTION_PROTOCOL_FILENAME  	methods.docx
CO:SAMPLE_TYPE                   	Urine
#TREATMENT
TR:TREATMENT_SUMMARY             	A spike-in experiment was conducted by adding a mixture of compounds to a urine
TR:TREATMENT_SUMMARY             	sample. 540 µL of a human urine standard material (Golden West Diagnostics,
TR:TREATMENT_SUMMARY             	LLC) was mixed with 60 µL of phosphate buffer (1.5 mol L-1 phosphate; 1.1 mmol
TR:TREATMENT_SUMMARY             	L-1 DSS; pH 7.4) following a previous procedure 1, and 20 µL of a spike-in
TR:TREATMENT_SUMMARY             	solution was sequentially added to the sample.
#SAMPLEPREP
SP:SAMPLEPREP_SUMMARY            	n/a
#ANALYSIS
AN:ANALYSIS_TYPE                 	NMR
AN:LABORATORY_NAME               	Arthur S. Edison Lab
AN:OPERATOR_NAME                 	Mario Uchimiya
AN:DETECTOR_TYPE                 	Avance III 600 MHz spectrometer equipped with a 5mm TCI cryoprobe
AN:ANALYSIS_PROTOCOL_FILE        	https://github.com/edisonomics/SAND/tree/main/scripts/spike_urine
AN:PROCESSING_PARAMETERS_FILE    	https://github.com/edisonomics/SAND/tree/main/scripts/spike_urine
AN:DATA_FORMAT                   	NMR
#NMR
NM:INSTRUMENT_NAME               	Avance III 600 MHz spectrometer equipped with a 5mm TCI cryoprobe
NM:INSTRUMENT_TYPE               	FT-NMR
NM:NMR_EXPERIMENT_TYPE           	1D-1H
NM:SPECTROMETER_FREQUENCY        	600 MHZ
NM:NMR_PROBE                     	cryoprobe
NM:TEMPERATURE                   	24.85
NM:NUMBER_OF_SCANS               	64
NM:RELAXATION_DELAY              	4s
NM:SPECTRAL_WIDTH                	16.00 ppm
NM:NUM_DATA_POINTS_ACQUIRED      	32K
NM:NMR_RESULTS_FILE              	quantification_NMR.txt UNITS:Amplitude
#END