#METABOLOMICS WORKBENCH wabdelmoula_20210620_173225 DATATRACK_ID:2703 STUDY_ID:ST001857 ANALYSIS_ID:AN003010 PROJECT_ID:PR001171 VERSION 1 CREATED_ON July 12, 2021, 9:24 pm #PROJECT PR:PROJECT_TITLE msiPL: Peak Learning of Mass Spectrometry Imaging Data Using Artificial Neural PR:PROJECT_TITLE Networks PR:PROJECT_TYPE Deep Learning for MSI data analysis PR:PROJECT_SUMMARY The attached MSI datasets of GBM and prostate cancer tissues were analyzed in PR:PROJECT_SUMMARY the manuscript by Abdelmoula et al. (bioRxiv 2020.08.13.250142). The below is PR:PROJECT_SUMMARY taken from the abstract: Mass spectrometry imaging (MSI) is an emerging PR:PROJECT_SUMMARY technology that holds potential for improving clinical diagnosis, biomarker PR:PROJECT_SUMMARY discovery, metabolomics research and pharmaceutical applications. The large data PR:PROJECT_SUMMARY size and high dimensional nature of MSI pose computational and memory PR:PROJECT_SUMMARY complexities that hinder accurate identification of biologically-relevant PR:PROJECT_SUMMARY molecular patterns. We propose msiPL, a robust and generic probabilistic PR:PROJECT_SUMMARY generative model based on a fully-connected variational autoencoder for PR:PROJECT_SUMMARY unsupervised analysis and peak learning of MSI data. The method can efficiently PR:PROJECT_SUMMARY learn and visualize the underlying non-linear spectral manifold, reveal PR:PROJECT_SUMMARY biologically-relevant clusters of tumor heterogeneity and identify underlying PR:PROJECT_SUMMARY informative m/z peaks. The method provides a probabilistic parametric mapping to PR:PROJECT_SUMMARY allow a trained model to rapidly analyze a new unseen MSI dataset in a few PR:PROJECT_SUMMARY seconds. The computational model features a memory-efficient implementation PR:PROJECT_SUMMARY using a minibatch processing strategy to enable the analyses of big MSI data PR:PROJECT_SUMMARY (encompassing more than 1 million high-dimensional datapoints) with PR:PROJECT_SUMMARY significantly less memory. We demonstrate the robustness and generic PR:PROJECT_SUMMARY applicability of the application on MSI data of large size from different PR:PROJECT_SUMMARY biological systems and acquired using different mass spectrometers at different PR:PROJECT_SUMMARY centers, namely: 2D Matrix-Assisted Laser Desorption Ionization (MALDI) Fourier PR:PROJECT_SUMMARY Transform Ion Cyclotron Resonance (FT ICR) MSI data of human prostate cancer, 3D PR:PROJECT_SUMMARY MALDI Time-of-Flight (TOF) MSI data of human oral squamous cell carcinoma, 3D PR:PROJECT_SUMMARY Desorption Electrospray Ionization (DESI) Orbitrap MSI data of human colorectal PR:PROJECT_SUMMARY adenocarcinoma, 3D MALDI TOF MSI data of mouse kidney, and 3D MALDI FT ICR MSI PR:PROJECT_SUMMARY data of a patient-derived xenograft (PDX) mouse brain model of glioblastoma. PR:INSTITUTE Brigham and Women's Hospital PR:DEPARTMENT Neurosurgery PR:LABORATORY Surgical Molecular Imaging Laboratory PR:LAST_NAME Abdelmoula PR:FIRST_NAME Walid PR:ADDRESS 60 Fenwood RD, Boston, Massachusetts, 02115, USA PR:EMAIL wahassan@bwh.harvard.edu PR:PHONE 617-525-7374 #STUDY ST:STUDY_TITLE Peak Learning of Mass Spectrometry Imaging Data Using Artificial Neural Networks ST:STUDY_TITLE (Prostate tissue) ST:STUDY_SUMMARY The human prostate tissue sample was analyzed by 2D MALDI FT ICR MSI. For ST:STUDY_SUMMARY detailed information we refer to the msiPL manuscript by Abdelmoula et al. ST:INSTITUTE Brigham and Women’s Hospital ST:LAST_NAME Abdelmoula ST:FIRST_NAME Walid ST:ADDRESS 60 Fenwood RD, Boston, MA ST:EMAIL wahassan@bwh.harvard.edu ST:PHONE 8572149765 #SUBJECT SU:SUBJECT_TYPE Human;Mouse SU:SUBJECT_SPECIES Homo sapiens; Mus musculus SU:TAXONOMY_ID 9606;10090 #FACTORS #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 - P_1900 Tissue Type:Human Prostate RAW_FILE_NAME=2D MALDI FT ICR MSI dataset SUBJECT_SAMPLE_FACTORS - Dataset_S1 Tissue Type:PDX GBM - mouse brain tumor section RAW_FILE_NAME=Dataset_S1.h5 SUBJECT_SAMPLE_FACTORS - Dataset_S2 Tissue Type:PDX GBM - mouse brain tumor section RAW_FILE_NAME=Dataset_S2.h5 SUBJECT_SAMPLE_FACTORS - Dataset_S3 Tissue Type:PDX GBM - mouse brain tumor section RAW_FILE_NAME=Dataset_S3.h5 SUBJECT_SAMPLE_FACTORS - Dataset_S4 Tissue Type:PDX GBM - mouse brain tumor section RAW_FILE_NAME=Dataset_S4.h5 #COLLECTION CO:COLLECTION_SUMMARY As stated in the msiPL paper: "Briefly, 12 µm thickness prostate sample CO:COLLECTION_SUMMARY diagnosed with a Gleason score of (3+4)=7 were mounted on a microscopy glass CO:COLLECTION_SUMMARY slide and coated with CHCA (5 mg/mL in 70/30 methanol/water with 0.1% CO:COLLECTION_SUMMARY trifluoroacetic acid V/V) using an automated sprayer (TM-Sprayer, HTX Imaging, CO:COLLECTION_SUMMARY Carrboro, NC). The analysis of the samples was performed on a 9.4 Tesla SolariX CO:COLLECTION_SUMMARY XR FT ICR mass spectrometer (Bruker Daltonics, Billerica, MA) using the MALDI CO:COLLECTION_SUMMARY source in positive ion mode in the mass range between 250-1000 m/z, with a CO:COLLECTION_SUMMARY spatial resolution of 120 µm." CO:SAMPLE_TYPE Prostate #TREATMENT TR:TREATMENT_SUMMARY N/A #SAMPLEPREP SP:SAMPLEPREP_SUMMARY As stated in the msiPL paper: For the Prostate tissue: "Briefly, 12 µm SP:SAMPLEPREP_SUMMARY thickness prostate sample diagnosed with a Gleason score of (3+4)=7 were mounted SP:SAMPLEPREP_SUMMARY on a microscopy glass slide and coated with CHCA (5 mg/mL in 70/30 SP:SAMPLEPREP_SUMMARY methanol/water with 0.1% trifluoroacetic acid V/V) using an automated sprayer SP:SAMPLEPREP_SUMMARY (TM-Sprayer, HTX Imaging, Carrboro, NC). The analysis of the samples was SP:SAMPLEPREP_SUMMARY performed on a 9.4 Tesla SolariX XR FT ICR mass spectrometer (Bruker Daltonics, SP:SAMPLEPREP_SUMMARY Billerica, MA) using the MALDI source in positive ion mode in the mass range SP:SAMPLEPREP_SUMMARY between 250-1000 m/z, with a spatial resolution of 120 µm." For the PDX GBM SP:SAMPLEPREP_SUMMARY mouse brain dataset: "The intracranial tumor belonging to a PDX model of GBM12 SP:SAMPLEPREP_SUMMARY (PDX National Resource, Mayo Clinic), was analyzed by MALDI FT ICR MSI using a SP:SAMPLEPREP_SUMMARY 9.4 Tesla SolariX mass spectrometer (Bruker Daltonics, Billerica, MA), using SP:SAMPLEPREP_SUMMARY continuous accumulation of selected ions in the mass range between 380-620 m/z. SP:SAMPLEPREP_SUMMARY The indium tin oxide (ITO)-coated slide with 12 µm thickness tissue sections , SP:SAMPLEPREP_SUMMARY was coated with DHB (160 mg/mL in a 70/30 v/v solution of methanol/0.2% TFA), SP:SAMPLEPREP_SUMMARY according to Randall et al2. The 3D MSI dataset was collected from 4 tissue SP:SAMPLEPREP_SUMMARY sections with an inter-slice distance of 160 µm. Internal online calibration SP:SAMPLEPREP_SUMMARY was performed using heme m/z 616.1776 during data acquisition." #CHROMATOGRAPHY CH:CHROMATOGRAPHY_TYPE Other CH:INSTRUMENT_NAME none CH:COLUMN_NAME none #ANALYSIS AN:ANALYSIS_TYPE MS #MS MS:INSTRUMENT_NAME Bruker Solarix FT-ICR-MS MS:INSTRUMENT_TYPE FT-ICR MS:MS_TYPE MALDI MS:ION_MODE POSITIVE MS:MS_COMMENTS Bruker software MS:MS_RESULTS_FILE ST001857_AN003010_Results.txt UNITS:Da Has m/z:Yes Has RT:No RT units:No RT data #END