Rapid and Accurate Identification of Microorganisms Using Ion Mobility–Mass Spectrometry

Abstract

Accurate identification of microorganisms to the strain and substrain levels in clinical and environmental samples is essential to provide an appropriate anti-biotherapy to the patients and reduce the prescription of broad-spectrum antimicrobials to minimize antibiotic resistance. Unfortunately, the current diagnosis methods are often slow, expensive, or laborious, which limits their use in resource-limited regions. Therefore, there is a strong unmet need for new technologies that can rapidly identify microorganisms in complex samples to complement the existing commercially available technologies. This Young Scientist Perspective demonstrates the value of combining the attributes of ion mobility-mass spectrometry and ambient ionization, enabling the rapid and accurate discrimination of bacteria to the species level after only a four-hour culturing period and showing that various bacterial species can have different isomers and conformers of their biomarkers. However, to discriminate closely-related bacterial strains, we needed to include other separation techniques in our workflow, such as liquid chromatography. Also, we utilized whole organism fingerprints, which include metabolites, lipids, and peptides, using our optimized workflow and machine learning to analyze a wide set of E. coli strains in artificially contaminated urine samples. Moreover, the various challenges for the routine identification of microorganisms using our optimized techniques in medical, environmental, and security fields and future outlooks are discussed.

Graphical Abstract

1. Introduction

1.1. Significance and current methods

Rapid and accurate identification of microorganisms is of increasing importance, with attention arising from public health concerns, environmental monitoring, and food safety surveillance [1–3]. The misuse and excessive use of antimicrobials lead to antimicrobial resistance, which is one of the top global public health and development threats; it could result in $1 trillion in additional healthcare costs by 2050 and over $1 trillion in gross domestic product (GDP) losses per year by 2030 [4]. Moreover, antimicrobial resistance directly led to 1.2 million deaths globally in 2019 [5]. If left unaddressed, there is a projection of 39 million deaths between now and 2050 due to antimicrobial resistance, highlighting the need for rapid and unequivocal identification of microorganisms [6]. Therefore, effective treatments would build on the rapid and accurate identification of the causative pathogens, which will not only lead to the prevention and successful treatment of various infections, leading to shorter hospitalization stays and enhanced food security, but also reduce the risk of developing antimicrobial resistance and improving overall cost-effectiveness [7, 8].

Conventional methods for the detection and identification of microorganisms based on culturing followed by standard biochemical identifications are selective, but they are often time-consuming and laborious [9, 10]. In addition, despite the efficiency of typical identification methods, such as enzyme-linked immunosorbent assay (ELISA), polymerase chain reaction (PCR), and pulsed-field gel electrophoresis (PFGE), they are labor-intensive, time-consuming, and expensive [11–13]. In fact, a combination of these methods for the detection of a particular pathogen is recommended when one detection method is not sufficient to confirm the detected pathogens [14].

More recently, matrix-assisted laser desorption ionization-mass spectrometry (MALDI-MS), a powerful tool in the clinical microbiology field, was approved for use in clinical microbiology laboratories and was reported to be efficiently utilized to identify microorganisms within minutes by comparing a characteristic spectrum of every species with a database. This allowed for a substantial reduction in the reagent and labor costs [15, 16]. In addition to its success in the detection and identification of various classes of microorganisms, MALDI-TOF MS has been used to detect microbial toxins and study antibiotic resistance [17–19]. Continuous upgrades of spectral databases and optimal sample enrichment will increase the power and potential applications of MALDI-TOF MS in clinical microbiology [18, 19].

1.2. New technologies and our role

To enable the direct analysis of microorganisms rapidly with minimal sample preparation requirements, several ambient ionization methods have been developed [20, 21]. Ambient ionization methods generate ions outside the vacuum of the mass spectrometer and require little or no sample preparation [22, 23]. Ambient ionization methods including desorption electrospray ionization (DESI) [24], low temperature plasma (LTP) [25], nanospray desorption electrospray ionization (nanoDESI) [26], liquid microjunction surface sampling probe (LMJ-SSP) [27], rapid evaporative ionization MS (REIMS) [28], laser ablation electrospray ionization (LAESI) [29], paper spray (PS) [30], touch spray (TS) [31], and direct analysis in real time (DART) [32] have been reported to be efficient in discriminating various microorganisms. My start in this research area was during my time as a postdoc at Purdue University working under the supervision of R. Graham Cooks, where the development of several ambient ionization techniques and their various applications were recently demonstrated. At Purdue University, I focused on developing PS and TS ambient ionization techniques to enable the species-level differentiation of wider sets of bacteria and fungi with high prediction rates requiring minimal sample preparation. Our initial studies were focused on developing and validating the use of PS-MS coupled with multivariate statistical tools, such as principal component analysis (PCA) and linear discriminant analysis (LDA) in the rapid discrimination of bacteria. These studies demonstrated that PS-MS can be used to efficiently generate characteristic and reproducible mass spectra directly from colonies grown on agar without requiring any sample preparation or lipid extraction steps prior to analysis. Our study included six Gram-positive bacteria species that have been differentiated with a prediction rate of 98% utilizing negative ions solely, while the sole utilization of positive ions has yielded a prediction rate of 96% [30]. When we tried to differentiate ten more closely related Gram-negative bacteria, lower prediction rates of 71% utilizing the negative ion mode’s information were obtained. Upon utilizing data fusion statistical methodology of the positive and negative ion modes’ chemical information, a prediction rate of 87% was achieved [30]. Whereas most mass spectrometry reports on the detection and differentiation of microorganisms have focused on bacteria, we utilized the PS-MS method to discriminate fungi rapidly with minimal sample preparation. Noteworthily, it was necessary to include a pretreatment step by adding a lysis solvent (e.g., N, N-dimethylformamide) to lyse the eukaryotic cell membrane of fungi. Then, they are subsequently analyzed by adding the spray solvent onto the paper triangle. In that study, eight closely related Candida species chosen for their clinical relevance and known taxonomy were investigated, and a prediction rate of 90% was achievable [33].

In my time as a postdoc in the laboratory of Richard D. Smith at Pacific Northwest National Laboratory and later as a senior scientist at MOBILion Systems, Inc., I actively worked on developing high-resolution ion mobility spectrometry (IMS) platforms, such as Structures for Lossless Ion Manipulations (SLIM) devices [34, 35]. Therefore, my first goals in my independent career were to establish innovative methods and instrumentation platforms, starting by coupling ambient ionization techniques to IMS-based instruments to enable the rapid diagnosis of infectious diseases and the efficient detection of various environmental contaminants and to provide insights into the structures of key biomarkers that can be used in diagnostics. In this Young Scientist Feature, unlike a traditional review, I will focus primarily on the contributions from our research laboratory in the field of the identification of microorganisms through IM-MS analysis.

2. Discrimination of bacteria using ambient ionization ion mobility-mass spectrometry

2.1. Optimization of the coupling of ambient ionization techniques with a commercial ion mobility mass spectrometer

While ambient ionization techniques enable the analysis of various samples in their native environments with little or no sample preparation within a few minutes, their reduced selectivity, even when coupled with high-resolution mass spectrometers, limits their application in certain fields. Since IMS is a gas phase separation technique in which structurally distinct compounds are distinguished based on collisions between analyte ions and inert buffer gas molecules under the influence of a weak electric field, they have been increasingly coupled with ambient ionization mass spectrometry to overcome challenges associated with the analysis of complex samples (e.g., the inability of MS to distinguish isomeric compounds with identical chemical formulae) [36–38]. Therefore, our first research project focused on coupling ambient ionization techniques, e.g., PS and leaf spray (LS), with a commercial ion mobility mass spectrometer. To this end, I acquired a drift tube ion mobility spectrometry (DTIMS) platform (Agilent 6560, Agilent Technologies, Santa Clara, CA) as it has the advantage that it can measure the collision cross section (CCS) of different ions directly without the need for external calibrants [39–41].

In our initial trials, we faced several challenges related to the design and typical parameters applied in the commercial Agilent 6560 instrument which uses an electrospray (ESI) ionization source that employs drying gas at high temperatures which leads to fast evaporation of the spray solvent and sample which in turn causes the ion signal to disappear after a few seconds. With the help of Shane Tichy (Agilent Technologies) and Brett Marsh (Purdue University), we tested various commercially available spray shields, which prevented the fast evaporation of the sample and spray solvent. We systematically optimized and validated our setup to perform PS–IM–MS and LS–IM–MS analyses efficiently for analyzing various constitutional and configurational isomers, resulting in high sensitivity and selectivity [42]. Moreover, we optimized several experimental parameters, including spray shield, distance from IM–MS inlet, capillary voltage (Vcap), and drying gas temperature and flow rate, as shown in Figure 1. We achieved higher performance when we used the vortex spray shield since it allowed the transmission of more ions while deflecting more drying gas away from the paper substrate, resulting in higher ion signals and longer spray times. This was attributed to the presence of side openings that divert the drying gas away from the paper triangle, unlike the other spray shields we examined. An ion signal was observed immediately upon the application of the spray solvent and voltage, and the signal disappeared reproducibly once the spray voltage was turned off, as shown in Figure 1A. As we varied the magnitude of the applied capillary voltage between −0.5 and −4.5 kV, the ion signal reached the highest intensity at −3.5 kV (Figure 1B). Moreover, a significant decrease in the ion signal was observed, similar to previous reports upon increasing the Vcap beyond the optimum voltage [43]. Moreover, we found that the optimal distance between the IM–MS inlet and the paper spray tip along the x-axis was ~6 mm (Figure 1C), while electrical discharge occurred at a distance of less than 2 mm, which was consistent with previous studies [37].

Figure 1.

Paper spray (PS) ambient ionization source optimization using S-(−)-verapamil in the positive ion mode under various conditions, which include: (A) with (On) and without (Off) applied voltage gradient. (B) various capillary voltages (Vcap). (C) different distances of PS tip from IM-MS inlet. (D) various drying gas temperatures where the drying gas flow rate was maintained at 5 L/min (top) and flow rates where the drying gas temperature was maintained at 250°C (bottom) on spray time (red triangles) and signal intensity (blue circles). Reprinted with permission from Journal of the American Society for Mass Spectrometry, 33 (1), Systematic Optimization of Ambient Ionization Ion Mobility-Mass Spectrometry for Rapid Separation of Isomers, 160–171, Copyright 2022, American Chemical Society.

In addition, the effects of varying the drying gas temperature and flow rate on the analyte’s ion intensity and spray time were examined (Figure 1D). In that study, the optimized drying gas temperature and the flow rate were 250 °C and 8 L/min, respectively, which was attributed to the increased evaporation of the solvent from the solvent droplets followed by the subsequent desolvation process that generates more gas-phase ions transmitted into the IM–MS inlet. Moreover, the spray time decreased when the drying gas temperature and flow rate increased due to the rapid evaporation of the sample by the drying gas.

The optimized PS-IM-MS and LS-IM-MS platforms were proven successful in the rapid separation of isomers; we applied them to the analysis and separation of isomeric pesticides (propazine and terbuthylazine) and lipids (cis and trans) in the positive and negative ion modes, respectively. Moreover, we used advanced computational tools to confidently identify analyte structures by comparing collision cross section (CCS) values from experimental IM measurements and theoretical calculations [44].

2.2. Discrimination of Microorganisms by High-Resolution Paper Spray – Ion Mobility – Mass Spectrometry

Although previous studies illustrated that the PS-MS technique was successful in discriminating various microorganisms utilizing their lipid and metabolic profiles, long incubation periods (~ 24–48 hours) were typically needed before a successful analysis, and poor separation and overlap between bacteria at the species level when discrimination was solely based on mass spectra [30, 33, 45, 46]. We attributed the inability of PS-MS methods to discriminate microorganisms with high prediction rates to the MS’s limited specificity, which limits its ability to differentiate biomarkers with identical chemical formulas or the same m/z. Therefore, we hypothesized that since key biomarkers, such as lipids and metabolites, can have several isomeric forms, revealing these differences can increase our capability in discriminating closely-related microorganisms. Therefore, in our first application study, we optimized and examined PS-IM-MS/MS methods to distinguish five Bacillus species rapidly and unambiguously to the species level directly in ~1–2 minute runs after only 4 hours of incubation time [47]. Noteworthily, surfactins and phospholipids were the major ion populations detected in our mass spectra.

To quantify the impact of including IM separations into the PS-MS workflows, we compared the prediction rates obtained using PS-MS/MS with those obtained by PS-IM-MS/MS. We found that upon including the IM separation dimension data into the PS-MS/MS workflow (i.e., PS-IM-MS/MS workflow), the prediction rates increased significantly in the negative and positive ion modes from 92.4% and 97.6% to 99.7% and 100.0%, respectively [47]. The increased prediction rates have been associated with the capability of IM to provide a clear separation between the diagnostic isomers, leading to enhanced species-level identification.

To explain the enhanced prediction rates achieved by PS-IM-MS/MS over PS-MS/MS analyses, we carefully investigated the drift time spectra of the 5 Bacillus species obtained from our PS-IM-MS studies. They revealed different isomeric forms of several phospholipids (e.g., PG (32:0) and PG (30:0)). Phospholipids were selected because they were reported previously to exist in various isomeric forms in Bacillus cell membranes, such as acyl chain isomers, positional isomers (e.g., sn- and double bond positional isomers), and stereoisomers (e.g., cis/trans isomers) [48, 49]. Also, phospholipids were detected in the 5 Bacillus species we investigated in contrast to other metabolites (e.g., surfactins and triglycerides) that were absent from some species.

The numbers of the resolved IM peaks (Figure 2A) observed for the PG (32:0) across the 5 Bacillus species were 3 for B. thuringiensis, 2 for B. subtilis and B. velezensis, and 1 for B. altitudinis and B. pumilus. Moreover, we confirmed the identities of the lipid isomers as PG (16:0/16:0), PG (15:0/17:0), and PG (14:0/18:0) by comparing the measured and predicted CCS values. Of note, the percentage difference between the measured and predicted values was less than 0.8%. To confirm the identities of these isomers, we used tandem mass spectrometry measurements, which revealed characteristic fragment ions of the PG (32:0) distinctively present in each species (Figure 2B). For instance, the CID spectrum of m/z 721.51 (PG (32:0)) in B. thuringiensis demonstrated the presence of the three acyl chain isomers identified by the IM separations and CCS measurements. Similarly, the CID spectrum of PG (32:0) in B. altitudinis and B. pumilus confirmed the presence of PG (15:0/17:0) isomer as accurately identified by IM. In addition, the MS/MS spectrum of PG (32:0) in B. subtilis and B. velezensis confirmed the presence of PG (15:0/17:0) and PG (16:0/16:0) isomers as accurately identified by the IM measurements. These results indicate that PG (15:0/17:0) was present in all the Bacillus species, PG (16:0/16:0) was present in three of the five Bacillus species, B. thuringiensis, B. subtilis, and B. velezensis, while PG (14:0/18:0) was unique to B. thuringiensis.

Figure 2.

PS-IM-MS and PS-MS/MS spectra obtained in the negative ion mode of the 5 Bacillus species where (A) Ion mobility spectra revealed the presence of various lipid isomers of m/z 721.51 (PG (32:0)) and (B) Tandem mass spectra (MS/MS) spectra of m/z 721.51 which support the identification of the lipid isomers by the IM spectra. Reprinted with permission from International Journal of Mass Spectrometry, 478, Species-level discrimination of microorganisms by high-resolution paper spray–Ion mobility–Mass spectrometry, 116871, Copyright 2022, Elsevier.

3. Strain-Level Discrimination of Microorganisms by High-Resolution Ion Mobility – Mass Spectrometry

Since different strains share a close genetic identity but can be associated with various diseases and antibiotic susceptibility profiles, we decided to challenge our previously optimized PS-IM-MS/MS workflow to discriminate seven non-pathogenic Escherichia coli (E. coli) strains (K-12, C41, BL21, CSH23, DH10B, DH5α, and S17–1 λpir) in the negative and positive ionization modes [50]. Interestingly, careful examination of the IM, MS, and MS/MS extracted ion spectra of several phospholipid ions (e.g., [PG (33:1)-H]⁻ and [PE (33:1)+Na]⁺) across the seven E. coli strains revealed the existence of various conformers and isomers for these phospholipids. Moreover, the identity of these lipids was confirmed by comparing their experimental CCS values with their corresponding theoretical ones. We obtained 62.5% and 73.5% prediction rates in the negative and positive ion modes, respectively, as quantified by LDA [50]. Fusing the chemical information present in the negative and positive ion data increased the separation among the seven E. coli strains, as visible in the PCA score plot (Figure 3A), resulting in a significantly increased prediction rate of 80.5% as quantified by LDA. Although these findings demonstrated that PS-IM-MS/MS can be an effective analytical technique for rapid detection and accurate discrimination of bacteria strains, higher prediction rates are desired.

Figure 3.

PCA plots of seven E. coli strains using (A) the PS-IM-MS/MS method in the Data Fusion mode of the extracted PCs of the negative and positive ion modes information and the LC-IM-MS/MS method in the (B) negative and (C) positive ion modes. The E. coli strains are indicated by color and shape as follows: BL21 (Black circles), C41(Red triangles pointing up), CSH23 (Blue triangles pointing down), DH10B (Green diamonds), DH5α (Magenta squares), K12 (Orange stars), and S17–1 λpir (Wine hexagons). Adapted with permission from Journal of the American Society for Mass Spectrometry, 34 (6), Strain-level discrimination of bacteria by liquid chromatography and paper spray ion mobility-mass spectrometry, 1125–1135, Copyright 2023, American Chemical Society.

The low predictive ability of the PS-IM-MS/MS has been attributed to the ionization suppression from matrix effects, the absence of chromatographic separation, and the limited number of key biomarkers. Therefore, we implemented multiple routes to increase the prediction capabilities of our methods, such as the modification of our workflow through the integration of orthogonal separation methods, including additional biomarkers into our analysis, and utilizing advanced multidimensional statistical analysis models, such as machine learning algorithms.

Analyzing lipid standards and E. coli lipid extracts demonstrated that integrating two orthogonal separation methods (LC and IM) with tandem mass analysis (i.e., LC-IM-MS/MS workflow) allowed multidimensional complementary separation and increased peak capacity, revealing the existence of more isomers and conformers that wouldn’t be observed with LC, IM or MS/MS alone. Therefore, we optimized and utilized an LC-IM-MS/MS workflow to analyze the seven E. coli strains. We found that isomeric lipid biomarkers discriminating E. coli strains were resolved in at least one separation dimension. The high selectivity of LC-IM-MS/MS was reflected in the PCA score plots, which showed a clear separation between E. coli strains in negative and positive ion modes (Figures 3B and 3C, respectively). The quantification of separation using LDA resulted in 96.1% and 100% prediction rates in the negative and positive ion modes, respectively, showing a high classification rate of these strains. Moreover, the bulk of our experimental analysis was supported by comparative genome analysis. Therefore, our results showed that multidimensional LC-IM-MS/MS is a highly selective and sensitive discrimination technique for closely-related bacterial strains.

Since our main goal is to identify microorganisms present in complex clinical and environmental samples rapidly and accurately, we further optimized and utilized LC-IM-MS/MS and machine learning algorithms to accurately identify and distinguish five pathogenic (O157, O104, O111, O18, and O16) and six non-pathogenic strains (K-12, C41, BL21, DH5α, DH10B, S17–1 λpir) E. coli strains in artificially contaminated urine samples [51]. In that study, we used whole organism fingerprinting, including metabolites, lipids, and peptides, to distinguish the eleven E. coli strains and create a spectral library of biomarkers for individual strains. Moreover, the unique list of metabolites, lipids, and peptides discovered by machine learning for each strain was then tested on artificially contaminated urine samples to examine bacteria strain identification and discrimination in urine samples.

The top 10 key features significantly contributing to the accurate differentiation of the E. coli strains, including lipid biomarkers from both ionization modes and peptide signatures from positive ionization mode, were combined to obtain the whole organism fingerprint of each strain (Figure 4). Noteworthily, these E. coli lipid and peptide biomarkers were completely absent in control urine samples, highlighting the uniqueness of these biomarkers to the E. coli strains. Moreover, the top 100 features for positive and negative ion modes of lipids and positive ion mode of peptides were used for hierarchical clustering of the E. coli strains (Figure 4), which distinguished certain patterns within the lipid and peptide abundance profiles. We also observed that each category of omics provided complementary information, resulting in efficient discrimination of a wide set of closely-related E. coli strains. We found that utilizing machine learning for statistical analysis rather than PCA enhanced the accuracy of statistical analysis as machine learning, in contrast to PCA, utilizes a supervised approach. Hence, these results have shown the great potential of using LC-IM-MS/MS and machine learning for accurate detection and discrimination of bacterial and fungal strains present in various environmental and clinical samples.

Figure 4.

Heatmaps of the top 10 significant features obtained from data collected in the negative ion mode and positive ion mode from lipid and peptide ions. The hierarchical clustering dendrogram based on the lipid and peptide ions is presented under the heatmap of the multi-omics and the intensity of each feature can be read by the color scale located at the bottom left. Reprinted with permission from Journal of the American Society for Mass Spectrometry, 35 (11), Discrimination of Common E. coli Strains in Urine by Liquid Chromatography–Ion Mobility–Tandem Mass Spectrometry and Machine Learning, 2706–2713, Copyright 2024, American Chemical Society.

4. Conclusions, Challenges, and Future Outlook

Our studies have reinforced the idea that combining the attributes of ion mobility-mass spectrometry and ambient ionization enables the analysis of clinical and environmental samples rapidly and accurately with minimal sample preparation efforts. In addition to the improved signal-to-noise ratio and the removal of matrix interference due to ion mobility separations, IM measurements also elucidate structures based on experimentally obtaining their corresponding CCS values and comparing them with those obtained using computational methods [44, 52]. Moreover, our studies have shown that PS-IM-MS/MS enables rapid and accurate discrimination of bacteria to species level after only a four-hour culturing period and that various bacterial species can have different isomers and conformers of their biomarkers. However, it was fairly challenging for the PS-IM-MS/MS method to discriminate closely-related bacterial strains, which was attributed to the ionization suppression that is typically observed with ambient ionization techniques. Upon including other separation techniques into our workflow (i.e., by performing LC-IM-MS/MS analyses), prediction rates of nearly 100% were achieved. To get closer to analyzing clinical samples, we implemented whole organism fingerprints, which included metabolites, lipids, and peptides using LC-IM-MS/MS and machine learning, which was proven efficient in discriminating eleven E. coli strains in artificially contaminated urine samples.

Challenges for routinely using the optimized LC-IM-MS/MS techniques in medical, environmental, and security fields to identify microorganisms include the need for building extensive databases that cover the key biomarkers, the need for culturing prior to the analysis, the need to investigate the impact of varying culture conditions, such as culturing temperature and media on the prediction rates, the sophistication of the current methods and instrumentation, and the purchase and maintenance costs of these instruments. However, the outlook is very bright, as many of these challenges are being actively addressed by several research teams, including ours, with innovative ambient ionization-IM-MS methods and portable IM spectrometers, fully elucidating the structures of the key biomarkers with reactions (e.g., epoxidation) and computational methods (e.g., molecular dynamics simulations), and building extensive databases that include CCS values for the key metabolite, peptide, and lipid biomarkers. Future investigations should also aim at developing methods to overcome matrix effects associated with complex environmental and clinical samples. Moreover, since the coupling of LC with IM-MS led to a significant increase in the prediction rates, coupling ambient ionization techniques with higher resolution IM platforms, such as atmospheric pressure drift tube ion mobility spectrometry (AP-DTIMS, Rp ~ 250) [53], trapped ion mobility spectrometry (TIMS, Rp ~ 400) [53], cyclic traveling wave ion mobility spectrometry (cTWIMS, Rp ~ 750) [54], and structures for lossless ion manipulations (SLIM, Rp ~ 1860) [34, 55] in particular, if miniaturized will increase their adoption by more researchers in various fields and make them more accessible and affordable to be implemented in resource-limited settings. In fact, we are currently optimizing and validating an ion mobility spectrometer based on the SLIM technology that we recently built for the analysis of microorganisms.

Moreover, our studies showed that utilizing a whole-omics approach enables the discrimination of a wider set of bacterial strains with the support of machine learning. Therefore, we are working on expanding this work by utilizing chemical reactions to locate double bonds, high-resolution IM separation to discriminate isomers, innovative extraction and sample cleanup methods, and computational approaches to investigate the role of peptide, protein, metabolite, and lipid biomarkers in the identification of microorganisms. These efforts, along with other research groups’ efforts, will enable the detection and identification of co-occurring microorganisms present in low concentrations, which will get us closer to performing efficient, affordable, and high-throughput analysis of real environmental and clinical samples.

Highlights:

• This Young Scientist Perspective demonstrates the value of combining the attributes of ion mobility-mass spectrometry and ambient ionization for rapid and accurate discrimination of bacteria

• Ion mobility mass spectrometry enables the identification of biomarkers’ isomers to the species and strains levels

• A wide set of closely-related bacterial strains in urine samples was discriminated utilizing a whole omics approach supported by machine learning

• Various challenges for the routine identification of microorganisms using IM-based techniques in medical, environmental, and security fields and future outlooks are discussed

Biography

Ahmed Hamid is an Assistant Professor in the Department of Chemistry and Biochemistry at Auburn University. He received his PhD degree in Physical Chemistry from Virginia Commonwealth University in 2012 under the supervision of Prof. Samy El-Shall working on gas-phase ion chemistry. He did postdoctoral research with Prof. Graham Cooks at Purdue University from 2012-2014 and at Pacific Northwest National Laboratory with Dr. Richard D. Smith from 2014-2017. Then, he became a Senior Scientist at MOBILion Systems, Inc. from 2017 to 2019 before joining Auburn University in August 2019. His research is focused on the development of novel mass spectrometry instruments, in particular those utilizing ion mobility separations for several applications in clinical and environmental research areas. He was awarded the R&D 100 award in 2017, the Maximizing Investigators’ Research Award in 2022, and was featured on the cover of the Journal of the American Society for Mass Spectrometry as one of the Emerging Investigators in 2024.

Data availability

No data was used for the research described in the article.