scholarly journals FATES: A Flexible Analysis Toolkit for the Exploration of Single Particle Mass Spectrometer Data

2016 ◽  
Author(s):  
Camille M. Sultana ◽  
Gavin Cornwell ◽  
Paul Rodriguez ◽  
Kimberly A. Prather
Keyword(s):  
Mass Spectrometer ◽  
Single Particle ◽  
Mass Spectra ◽  
Large Datasets ◽  
Particle Mass ◽  
Data Exploration ◽  
Data Sets ◽  
Data Formats ◽  
Peak Intensity ◽  
Spectrometer Data

Abstract. Single particle mass spectrometer (SPMS) analysis of aerosols has become increasingly popular since its invention in the 1990s. Today many iterations of commercial and lab-built SPMS are in use worldwide. However supporting analysis toolkits for these powerful instruments are either outdated, have limited functionality, or are versions that are not available to the scientific community at large. In an effort to advance this field and allow better communication and collaboration between scientists we have developed FATES (Flexible Analysis Toolkit for the Exploration of SPMS data), a MATLAB toolkit easily extensible to an array of SPMS designs and data formats. FATES was developed to minimize the computational demands of working with large datasets while still allowing easy maintenance, modification, and utilization by novice programmers. FATES permits scientists to explore, without constraint, complex SPMS data with simple scripts in a language popular for scientific numerical analysis. In addition FATES contains an array of data visualization GUIs which can aid both novice and expert users in calibration of raw data, exploration of the dependence of mass spectra characteristics on size, time, and peak intensity, as well investigations of clustered data sets.

scholarly journals FATES: a flexible analysis toolkit for the exploration of single-particle mass spectrometer data

2017 ◽  
Vol 10 (4) ◽  
pp. 1323-1334 ◽  
Author(s):  
Camille M. Sultana ◽  
Gavin C. Cornwell ◽  
Paul Rodriguez ◽  
Kimberly A. Prather
Keyword(s):  
Mass Spectrometer ◽  
User Interfaces ◽  
Single Particle ◽  
Spectral Characteristics ◽  
Large Data ◽  
Particle Mass ◽  
Data Sets ◽  
Mass Spectral ◽  
Data Formats ◽  
Spectrometer Data

Abstract. Single-particle mass spectrometer (SPMS) analysis of aerosols has become increasingly popular since its invention in the 1990s. Today many iterations of commercial and lab-built SPMSs are in use worldwide. However, supporting analysis toolkits for these powerful instruments are outdated, have limited functionality, or are versions that are not available to the scientific community at large. In an effort to advance this field and allow better communication and collaboration between scientists, we have developed FATES (Flexible Analysis Toolkit for the Exploration of SPMS data), a MATLAB toolkit easily extensible to an array of SPMS designs and data formats. FATES was developed to minimize the computational demands of working with large data sets while still allowing easy maintenance, modification, and utilization by novice programmers. FATES permits scientists to explore, without constraint, complex SPMS data with simple scripts in a language popular for scientific numerical analysis. In addition FATES contains an array of data visualization graphic user interfaces (GUIs) which can aid both novice and expert users in calibration of raw data; exploration of the dependence of mass spectral characteristics on size, time, and peak intensity; and investigations of clustered data sets.

scholarly journals Development and characterization of a single particle laser ablation mass spectrometer (SPLAM) for organic aerosol studies

2012 ◽  
Vol 5 (1) ◽  
pp. 225-241 ◽  
Author(s):  
F. Gaie-Levrel ◽  
S. Perrier ◽  
E. Perraudin ◽  
C. Stoll ◽  
N. Grand ◽  
...  
Keyword(s):  
Laser Ablation ◽  
Detection Limit ◽  
Mass Spectrometer ◽  
Single Particle ◽  
Mass Spectra ◽  
Diode Lasers ◽  
Optical Detection ◽  
Organic Aerosol ◽  
Polystyrene Latex ◽  
Particle Mass

Abstract. A single particle instrument was developed for real-time analysis of organic aerosol. This instrument, named Single Particle Laser Ablation Mass Spectrometry (SPLAM), samples particles using an aerodynamic lens system for which the theoretical performances were calculated. At the outlet of this system, particle detection and sizing are realized by using two continuous diode lasers operating at λ = 403 nm. Polystyrene Latex (PSL), sodium chloride (NaCl) and dioctylphtalate (DOP) particles were used to characterize and calibrate optical detection of SPLAM. The optical detection limit (DL) and detection efficiency (DE) were determined using size-selected DOP particles. The DE ranges from 0.1 to 90% for 100 and 350 nm DOP particles respectively and the SPLAM instrument is able to detect and size-resolve particles as small as 110–120 nm. During optical detection, particle scattered light from the two diode lasers, is detected by two photomultipliers and the detected signals are used to trigger UV excimer laser (λ = 248 nm) used for one-step laser desorption ionization (LDI) of individual aerosol particles. The formed ions are analyzed by a 1 m linear time-of-flight mass spectrometer in order to access to the chemical composition of individual particles. The TOF-MS detection limit for gaseous aromatic compounds was determined to be 0.85 × 10−15 kg (∼4 × 103 molecules). DOP particles were also used to test the overall operation of the instrument. The analysis of a secondary organic aerosol, formed in a smog chamber by the ozonolysis of indene, is presented as a first application of the instrument. Single particle mass spectra were obtained with an effective hit rate of 8%. Some of these mass spectra were found to be very different from one particle to another possibly reflecting chemical differences within the investigated indene SOA particles. Our study shows that an exhaustive statistical analysis, over hundreds of particles, and adapted reference mass spectra are further needed to understand the chemical meaning of single particle mass spectra of chemically complex submicrometer-sized organic aerosols.

scholarly journals Organic particle types by single-particle measurements using a time-of-flight aerosol mass spectrometer coupled with a light scattering module

2012 ◽  
Vol 5 (2) ◽  
pp. 3047-3077 ◽  
Author(s):  
S. Liu ◽  
L. M. Russell ◽  
D. T. Sueper ◽  
T. B. Onasch
Keyword(s):  
Light Scattering ◽  
Mass Spectrometer ◽  
Single Particle ◽  
Mass Spectra ◽  
Time Of Flight ◽  
Particle Mass ◽  
Aerosol Mass Spectrometer ◽  
Particle Measurements ◽  
Aerosol Mass ◽  
Organic Particle

Abstract. Chemical and physical properties of individual ambient aerosol particles can vary greatly, so measuring the chemical composition at the single-particle level is essential for understanding atmospheric sources and transformations. Here we describe 46 days of single-particle measurements of atmospheric particles using a time-of-flight aerosol mass spectrometer coupled with a light scattering module (LS-ToF-AMS). The light scattering module optically detects particles larger than 180 nm vacuum aerodynamic diameter (130 nm geometric diameter) (with size resolution of 5–10 defined as dΔd at full width at half maximum) before they arrive at the chemical mass detector and then triggers the saving of single-particle mass spectra. 271 641 particles were detected and sampled during 237 h of sampling in single particle mode. By comparing the timing of light scattering and chemical ion signals for each particle, particle types were classified and their number fractions determined as follows: prompt vaporization (49%), delayed vaporization (7%), and null (44%). LS-ToF-AMS provided the first direct measurement of the size-resolved collection efficiency (CE) of ambient particles, with an approximate 50% number-based CE for particles above detection limit. Prompt and delayed vaporization particles (147 357 particles) were clustered based on similar organic mass spectra (using K-means algorithm) to result in three major clusters: highly oxidized particles (dominated by m/z 44), relatively less oxidized particles (dominated by m/z 43), and particles associated with fresh urban emissions. Each of the three organic clusters had limited chemical properties of other clusters, suggesting that all of the sampled organic particle types were internally mixed to some degree; however, the internal mixing was never uniform and distinct particle types existed throughout the study. Furthermore, the single particle mass spectra and diurnal variations of these clusters agreed well with mass-based components identified (using factor analysis) from simultaneous ensemble-averaged measurements, supporting the connection between ensemble-based factors and atmospheric particle sources and processes. Measurements in this study illustrate that LS-ToF-AMS provides unique information about organic particle types by number as well as mass.

scholarly journals Organic particle types by single-particle measurements using a time-of-flight aerosol mass spectrometer coupled with a light scattering module

2013 ◽  
Vol 6 (2) ◽  
pp. 187-197 ◽  
Author(s):  
S. Liu ◽  
L. M. Russell ◽  
D. T. Sueper ◽  
T. B. Onasch
Keyword(s):  
Light Scattering ◽  
Mass Spectrometer ◽  
Single Particle ◽  
Mass Spectra ◽  
Time Of Flight ◽  
Particle Mass ◽  
Aerosol Mass Spectrometer ◽  
Particle Measurements ◽  
Aerosol Mass ◽  
Organic Particle

Abstract. Chemical and physical properties of individual ambient aerosol particles can vary greatly, so measuring the chemical composition at the single-particle level is essential for understanding atmospheric sources and transformations. Here we describe 46 days of single-particle measurements of atmospheric particles using a time-of-flight aerosol mass spectrometer coupled with a light scattering module (LS-ToF-AMS). The light scattering module optically detects particles larger than 180 nm vacuum aerodynamic diameter (130 nm geometric diameter) before they arrive at the chemical mass spectrometer and then triggers the saving of single-particle mass spectra. 271 641 particles were detected and sampled during 237 h of sampling in single-particle mode. By comparing timing of the predicted chemical ion signals from the light scattering measurement with the measured chemical ion signals by the mass spectrometer for each particle, particle types were classified and their number fractions determined as follows: prompt vaporization (46%), delayed vaporization (6%), and null (48%), where null was operationally defined as less than 6 ions per particle. Prompt and delayed vaporization particles with sufficient chemical information (i.e., more than 40 ions per particle) were clustered based on similarity of organic mass spectra (using k-means algorithm) to result in three major clusters: highly oxidized particles (dominated by m/z 44), relatively less oxidized particles (dominated by m/z 43), and particles associated with fresh urban emissions. Each of the three organic clusters had limited chemical properties of other clusters, suggesting that all of the sampled organic particle types were internally mixed to some degree; however, the internal mixing was never uniform and distinct particle types existed throughout the study. Furthermore, the single-particle mass spectra and time series of these clusters agreed well with mass-based components identified (using factor analysis) from simultaneous ensemble-averaged measurements, supporting the connection between ensemble-based factors and atmospheric particle sources and processes. Measurements in this study illustrate that LS-ToF-AMS provides unique information about organic particle types by number as well as mass.

scholarly journals Sources and mixing state of size-resolved elemental carbon particles in a European megacity: Paris

2012 ◽  
Vol 12 (4) ◽  
pp. 1681-1700 ◽  
Author(s):  
R. M. Healy ◽  
J. Sciare ◽  
L. Poulain ◽  
K. Kamili ◽  
M. Merkel ◽  
...  
Keyword(s):  
Mass Spectrometer ◽  
Biomass Burning ◽  
Single Particle ◽  
Mass Concentration ◽  
Elemental Carbon ◽  
Fossil Fuel ◽  
Time Of Flight ◽  
Particle Mass ◽  
Mass Size ◽  
Mass Concentrations

Abstract. An Aerosol Time-Of-Flight Mass Spectrometer (ATOFMS) was deployed to investigate the size-resolved chemical composition of single particles at an urban background site in Paris, France, as part of the MEGAPOLI winter campaign in January/February 2010. ATOFMS particle counts were scaled to match coincident Twin Differential Mobility Particle Sizer (TDMPS) data in order to generate hourly size-resolved mass concentrations for the single particle classes observed. The total scaled ATOFMS particle mass concentration in the size range 150–1067 nm was found to agree very well with the sum of concurrent High-Resolution Time-of-Flight Aerosol Mass Spectrometer (HR-ToF-AMS) and Multi-Angle Absorption Photometer (MAAP) mass concentration measurements of organic carbon (OC), inorganic ions and black carbon (BC) (R2 = 0.91). Clustering analysis of the ATOFMS single particle mass spectra allowed the separation of elemental carbon (EC) particles into four classes: (i) EC attributed to biomass burning (ECbiomass), (ii) EC attributed to traffic (ECtraffic), (iii) EC internally mixed with OC and ammonium sulfate (ECOCSOx), and (iv) EC internally mixed with OC and ammonium nitrate (ECOCNOx). Average hourly mass concentrations for EC-containing particles detected by the ATOFMS were found to agree reasonably well with semi-continuous quantitative thermal/optical EC and optical BC measurements (r2 = 0.61 and 0.65–0.68 respectively, n = 552). The EC particle mass assigned to fossil fuel and biomass burning sources also agreed reasonably well with BC mass fractions assigned to the same sources using seven-wavelength aethalometer data (r2 = 0.60 and 0.48, respectively, n = 568). Agreement between the ATOFMS and other instrumentation improved noticeably when a period influenced by significantly aged, internally mixed EC particles was removed from the intercomparison. 88% and 12% of EC particle mass was apportioned to fossil fuel and biomass burning respectively using the ATOFMS data compared with 85% and 15% respectively for BC estimated from the aethalometer model. On average, the mass size distribution for EC particles is bimodal; the smaller mode is attributed to locally emitted, mostly externally mixed EC particles, while the larger mode is dominated by aged, internally mixed ECOCNOx particles associated with continental transport events. Periods of continental influence were identified using the Lagrangian Particle Dispersion Model (LPDM) "FLEXPART". A consistent minimum between the two EC mass size modes was observed at approximately 400 nm for the measurement period. EC particles below this size are attributed to local emissions using chemical mixing state information and contribute 79% of the scaled ATOFMS EC particle mass, while particles above this size are attributed to continental transport events and contribute 21% of the EC particle mass. These results clearly demonstrate the potential benefit of monitoring size-resolved mass concentrations for the separation of local and continental EC emissions. Knowledge of the relative input of these emissions is essential for assessing the effectiveness of local abatement strategies.

Bio-informatics and psychiatric epidemiology

2020 ◽  
pp. 359-372
Author(s):  
Nicola Voyle ◽  
Maximilian Kerz ◽  
Steven Kiddle ◽  
Richard Dobson
Keyword(s):  
Machine Learning ◽  
Data Cleaning ◽  
Epidemiological Studies ◽  
Selection Method ◽  
Psychiatric Epidemiology ◽  
Large Datasets ◽  
Data Exploration ◽  
Feature Identification ◽  
Data Formats ◽  
Method Selection

This chapter highlights the methodologies which are increasingly being applied to large datasets or ‘big data’, with an emphasis on bio-informatics. The first stage of any analysis is to collect data from a well-designed study. The chapter begins by looking at the raw data that arises from epidemiological studies and highlighting the first stages in creating clean data that can be used to draw informative conclusions through analysis. The remainder of the chapter covers data formats, data exploration, data cleaning, missing data (i.e. the lack of data for a variable in an observation), reproducibility, classification versus regression, feature identification and selection, method selection (e.g. supervised versus unsupervised machine learning), training a classifier, and drawing conclusions from modelling.

Sign in / Sign up

Export Citation Format

Share Document