Data Alignment Research Articles

New GO-based measures in multiple network alignment.

Protein-protein interaction (PPI) networks provide valuable insights into the function of biological systems. Aligning multiple PPI networks may expose relationships beyond those observable by pairwise comparisons. However, assessing the biological quality of multiple network alignments is a challenging problem. We propose two new measures to evaluate the quality of multiple network alignments using functional information from Gene Ontology (GO) terms. When aligning multiple real PPI networks across species, we observe that both measures are highly correlated with objective quality indicators, such as common orthologs. Additionally, our measures strongly correlate with an alignment's ability to predict novel GO annotations, which is a unique advantage over existing GO-based measures. The scripts and the links to the raw and alignment data can be accessed at https://github.com/kimiayazdani/GO_Measures.git.

Optimization of headspace high-capacity tool coupled to two-dimensional gas chromatography–mass spectrometry for mapping the volatile organic compounds of raw pistachios. A proof-of-concept on the classification ability by geographic origin

An optimized procedure for extracting and analyzing raw pistachio volatiles was developed through headspace sampling with high-capacity tools and subsequent analysis using comprehensive two-dimensional gas chromatography coupled with mass spectrometry. The examination of 18 pistachio samples belonging to different geographic areas led to the identification of a set of 99 volatile organic compounds (VOCs). Molecules were putatively identified using linear retention index, mass spectra similarity, and two-dimensional plot location. The impact of preprocessing and processing techniques on the aligned data matrix from a set of samples of different geographical origins, after removing contaminants, was evaluated. The combination of scaling with log-transformation, normalization with z-score, and data reduction with random forest machine learning algorithm generated a panel of 16 discriminatory VOC molecules. As a proof of concept, raw pistachios' VOC profile was employed for the first time to tentatively classify them based on their geographical origin.

The Condition of Education: Teacher Mental Health Absenteeism in the Aftermath of COVID-19

<i>The implications of teacher absenteeism are incommensurable. While absenteeism is a longstanding issue in education, there is a gap in the current literature addressing teachers’ mental health and absenteeism after schools reopened following the first wave of the COVID-19 virus. This qualitative descriptive study is anchored on the Job Demands-Job Resources (JD-R) model of organizational stress and well-being. The purpose of the study was twofold: a) to describe teacher absenteeism related to teachers’ mental health stemming from the pandemic’s stressful job demands after schools reopened in the school year 2021-2022 in two Eastern regions of the U.S., and b) to describe the availability of job resources such as administrative mental health support after schools reopened in the school year 2021-2022 in two Eastern regions of the U.S. The study sample consisted of 12 educators: Four teachers and eight administrators were selected using a purposive sample strategy. Semi-structured interviews were carefully designed to extract relevant data in alignment with the research questions. Thematic data analysis was conducted based on the study’s method and design. The study results revealed four themes: New and strenuous job demands, teachers’ mental health decline, mental health absenteeism, and availability of mental health administrative support to mitigate mental health absenteeism.</i>

Experimental investigation on a solar dryer assisted with minimum phase change material (PCM) placed on the inner walls of drying chamber

Solar dryers assisted with phase change material (PCM) speed up the drying process by compensating the negating effect of reduced or absence of solar energy under cloudy or non-sunshine hours respectively. The study of the relevant and recent literature indicated that the minimum quantity and proper placement of this PCM have not been explored to take full advantage of using them in solar dryers. Hence in this research work, experimental investigations on banana slice drying were carried out with 4 and 10 numbers of copper tubes filled with 176 and 494 kg of commercial paraffin wax respectively (correspondingly designated as SD-PCM1 and SD-PCM2). The copper tubes were placed in vertical and horizontal orientations on the inner wall of the drying chamber so that the upward movement of air would not disturbed. A comparison of the drying performance of a PCM-based unit with a natural convective solar dryer without PCM (SD) was also presented. As no appreciable differences between the mass loss during drying under SD and SD-PCM1 modes were observed, it was concluded that a mass of PCM greater than 176 g would enhance drying rates. Thus for the same final mass of dried banana slices, the drying time was about 15 and 9 h with SD and SD-PCM2 respectively. As the operating conditions of SD and SD-PCM vary considerably in terms of the thermal energy supplied during the drying process, the concept of a generalized drying curve with dimensionless time was used to study the drying kinetics. A close alignment of experimental data of moisture content ratio with the generalized drying curve under SD and SD-PCM modes indicates similar dryer performance irrespective of the differences in thermal energy supplied. The average exergetic efficiency of SD, SD-PCM1 and SD-PCM2 was 21, 26 and 28 % respectively. The sustainability index and environmental impact factor for SD-PCM2 were 1.41 and 2.5 respectively which was better compared to the respective values of 1.27 and 3.77 in SD mode favoring the use of PCM in solar dryers.

Humanitarian supply chain resilience: does organizational flexibility matter?

PurposeThe COVID-19 epidemic caused significant disruptions to numerous supply chains. In order to enhance the resilience of supply chains, Collaboration (CO), Information Alignment (IA), and Big Data Analytics Capability (BDAC) have emerged as contemporary strategies within the humanitarian context. This study was conducted to explore the mechanism via which the effect of BDAC, IA and CO on Humanitarian Supply Chain Resilience (HSCR) in the humanitarian space could be optimized through Organizational Flexibility (OF).Design/methodology/approachA model of six hypotheses was developed based on the Organizational Information Processing Theory (OIPT). Data from 127 supply chain managers in humanitarian organizations were used to test the hypotheses. The analysis employed both descriptive and inferential statistics using SPSS version 26 and Smart-PLS version 3.FindingsThe study revealed that BDAC, IA, and CO individually influence supply chain resilience in the humanitarian setting while OF did not moderate the relationship between BDAC, IA, CO, and HSCR.Practical implicationsIt is essential that humanitarian stakeholders prioritize factors that could increase supply chain resilience by employing contemporary BDA technologies, effective information flow, and collaborative strategies to set up a robust humanitarian SC system that could help lessen the impact of disasters.Originality/valueThis presents interesting insights that advance theoretical debates on how CO, IA, and BDAC under varying levels of OF could influence SCR in the humanitarian context. The paper further offers some useful guidance to managers in relief organizations who desire to build resilient supply chains by leveraging BDAC, collaboration and information alignment. Finally, the paper may also provoke future humanitarian scholars to replicate the study using different approaches.

Advanced Multimodal Sentiment Analysis with Enhanced Contextual Fusion and Robustness (AMSA-ECFR): Symmetry in Feature Integration and Data Alignment

Multimodal sentiment analysis, a significant challenge in artificial intelligence, necessitates the integration of various data modalities for accurate human emotion interpretation. This study introduces the Advanced Multimodal Sentiment Analysis with Enhanced Contextual Fusion and Robustness (AMSA-ECFR) framework, addressing the critical challenge of data sparsity in multimodal sentiment analysis. The main components of the proposed approach include a Transformer-based model employing BERT for deep semantic analysis of textual data, coupled with a Long Short-Term Memory (LSTM) network for encoding temporal acoustic features. Innovations in AMSA-ECFR encompass advanced feature encoding for temporal dynamics and an adaptive attention-based model for efficient cross-modal integration, achieving symmetry in the fusion and alignment of asynchronous multimodal data streams. Additionally, the framework employs generative models for intelligent approximation of missing features. It ensures robust alignment of high-level features with multimodal data context, effectively tackling issues of incomplete or noisy inputs. In simulation studies, the AMSA-ECFR model demonstrated superior performance against existing approaches. The symmetrical approach to feature integration and data alignment contributed significantly to the model’s robustness and precision. In simulations, the AMSA-ECFR model demonstrated a 10% higher accuracy and a 15% lower mean absolute error than the current best multimodal sentiment analysis frameworks.

Performance Evaluation of an EI&CI Dual Ionization TOFMS Hyphenated with a Flow Modulated GC×GC System.

The use and compatibility of a dual-ionization TOFMS operating an EI source and a CI source in parallel using a single TOF mass analyzer with flow modulated two-dimensional GC (GC×GC) is described. Important figures of merit of the mass spectrometer that are required for two-dimensional GC hyphenation such as acquisition speed, ion source response, EI/CI switching, the GC transfer, and data alignment are carefully investigated and addressed. Improved fast switching ion optics allow switching in a 100 Hz frequency between EI and CI spectra sampled from the same GC×GC effluent. The spectra quality also influenced by the preseparation, especially of the EI source, is compared to a standard setup operating a single quadrupole MS coupled to the same GC system. Further, two setups including and excluding an additional flame ionization detector are presented. High increments in CI sensitivities are achieved by utilizing the high pumping efficiencies of the CI stage of the used mass spectrometer. By leading high flow ratios of the GC×GC modulation flows toward the CI source, the intensity can be increased by factors of up to 37 while maintaining the pressure balance of the less robust EI source. Finally, thermal desorption GC×GC-EI&CI-TOFMS analyses of traffic emission samples from a federal highway in Germany are executed with the presented setup.

Highly sensitive two-dimensional profiling of N-linked glycans by hydrophilic interaction liquid chromatography and dual stacking capillary gel electrophoresis

BackgroundN-Glycosylation is one of the most important post-translational modifications in proteins. As the N-glycan profiles in biological samples are diverse and change according to the pathological condition, various profiling methods have been developed, such as liquid chromatography (LC), capillary electrophoresis (CE), and mass spectrometry. However, conventional analytical methods have limitations in sensitivity and/or resolution, hindering the discovery of minor but specific N-glycans that are important both in the basic glycobiology research and in the medical application as biomarkers. Therefore, a highly sensitive and high-resolution N-glycan profiling method is required. ResultsIn this study, we developed a novel two-dimensional (2D) separation system, which couples hydrophilic interaction liquid chromatography (HILIC) with capillary gel electrophoresis (CGE) via large-volume dual preconcentration by isotachophoresis and stacking (LDIS). Owing to the efficient preconcentration efficiency of LDIS, limit of detection reached 12 pM (60 amol, S/N = 3) with good calibration curve linearity (R2 > 0.999) in the 2D analysis of maltoheptaose. Finally, 2D profiling of N-glycans obtained from standard glycoproteins and cell lysates were demonstrated. High-resolution 2D profiles were successfully obtained by data alignment using triple internal standards. N-glycans were well distributed on the HILIC/CGE 2D plane based on the glycan size, number of sialic acids, linkage type, and so on. As a result, specific minor glycans were successfully identified in HepG2 and HeLa cell lysates. Significance and noveltyIn conclusion, the HILIC/CGE 2D analysis method showed sufficient sensitivity and resolution for identifying minor but specific N-glycans from complicated cellular samples, indicating the potential as a next-generation N-glycomics tool. Our novel approach for coupling LC and CE can also dramatically improve the sensitivity in other separation modes, which can be a new standard of 2D bioanalysis applicable not only to glycans, but also to other diverse biomolecules such as metabolites, proteins, and nucleic acids.

SANTO: a coarse-to-fine alignment and stitching method for spatial omics

With the flourishing of spatial omics technologies, alignment and stitching of slices becomes indispensable to decipher a holistic view of 3D molecular profile. However, existing alignment and stitching methods are unpractical to process large-scale and image-based spatial omics dataset due to extreme time consumption and unsatisfactory accuracy. Here we propose SANTO, a coarse-to-fine method targeting alignment and stitching tasks for spatial omics. SANTO firstly rapidly supplies reasonable spatial positions of two slices and identifies the overlap region. Then, SANTO refines the positions of two slices by considering spatial and omics patterns. Comprehensive experiments demonstrate the superior performance of SANTO over existing methods. Specifically, SANTO stitches cross-platform slices for breast cancer samples, enabling integration of complementary features to synergistically explore tumor microenvironment. SANTO is then applied to 3D-to-3D spatiotemporal alignment to study development of mouse embryo. Furthermore, SANTO enables cross-modality alignment of spatial transcriptomic and epigenomic data to understand complementary interactions.

Integrating anatomical and functional landmarks for interparticipant alignment of imaging data

Abstract Aligning brain maps using functional features rather than anatomical landmarks potentially improves individual identifiability and increases power in group neuroimaging studies. However, alignment based purely on functional magnetic resonance imaging (fMRI) risks omitting useful anatomical constraints. An optimized combination of anatomical and functional feature alignment could balance the advantages of each approach. We used 3T fMRI data from 80 Human Connectome Project participants during seven tasks. The effectiveness of functional and anatomical alignment methods was evaluated using inter-participant decoding accuracy. Functional alignment mapped vertices from participants to a template, aligning their fMRI responses to shared responses during movie viewing. The template was derived from the combined fMRI responses of a set of participants. We benchmarked the results against existing functional alignment methods including the Procrustes method and ridge regression. A common practice in the field is to use the same participants for the alignment cohort and for template generation. We found that this inflates decoding accuracies by mixing anatomical and functional alignment. Based on this, we recommend that a template’s generalizability should be evaluated against held-out participants. Building on these findings, we investigated whether inter-subject alignment could be improved by integrating anatomical and functional information. We studied a modified alignment method where a single parameter interpolates between pure functional alignment and anatomical alignment. Optimising the parameter with nested cross validation, we found that integrating anatomical and functional information robustly reduced noise and improved alignment across a variety of alignment methods. Combining anatomical and functional information accounts for individual heterogeneity in functional topographies while incorporating anatomical constraints. The integrated alignment described here improves inter-subject decoding using functional brain maps. These findings also demonstrate that brain anatomy provides a lens into the inherent variability of individual neural landscapes.

Improved characterization of single-cell RNA-seq libraries with paired-end avidity sequencing.

Prevailing poly(dT)-primed 3' single-cell RNA-seq protocols generate barcoded cDNA fragments containing the reverse transcriptase priming site, which is expected to be the poly(A) tail or a genomic adenine homopolymer. Direct sequencing across this priming site was historically difficult because of DNA sequencing errors induced by the homopolymeric primer at the 'barcode' end. Here, we evaluate the capability of "avidity base chemistry" DNA sequencing from Element Biosciences to sequence through this homopolymer accurately, and the impact of the additional cDNA sequence on read alignment and precise quantification of polyadenylation site usage. We find that the Element Aviti instrument sequences through the thymine homopolymer into the subsequent cDNA sequence without detectable loss of accuracy. The resulting paired-end alignments enable direct and independent assignment of reads to polyadenylation sites, which bypasses complexities and limitations of conventional approaches but does not consistently improve read mapping rates compared to single-end alignment. We also characterize low-level artifacts and arrive at an adjusted adapter trimming and alignment workflow that significantly improves the alignment of sequence data from Element and Illumina, particularly in the context of extended read lengths. Our analyses confirm that Element avidity sequencing is an effective alternative to Illumina sequencing for standard single-cell RNA-seq, particularly for polyadenylation site analyses but do not rule out the potential for similar performance from other emerging platforms.

Driving Reality vs. Simulator: Data Distinctions

As the automotive industry undergoes a phase of rapid transformation driven by technological advancements, the integration of driving simulators stands out as an important tool for research and development. The usage of such simulators offers a controlled environment for studying driver behavior; the alignment of data, however, remains a complex aspect that warrants a thorough investigation. This research investigates driver state classification using a dataset obtained from real-road and simulated conditions, recorded through JINS MEME ES_R smart glasses. The dataset encompasses electrooculography signals, with a focus on standardizing and processing the data for subsequent analysis. For this purpose, we used a recurrent neural network model, which yielded a high accuracy on the testing dataset (86.5%). The findings of this study indicate that the proposed methodology could be used in real scenarios and that it could be used for the development of intelligent transportation systems and driver monitoring technology.

Dynamics of Saltwater Intrusion in a Heterogeneous Coastal Environment: Experimental, DC Resistivity, and Numerical Modeling Approaches

Saltwater intrusion (SWI) is a critical concern affecting coastal groundwater sources due to natural and anthropogenic activities. The health of coastal aquifers is deteriorated by excessive SWI, mainly caused by the disturbance of the freshwater–saltwater equilibrium due to the escalating population, climate change, and the rising demand for freshwater resources for human activities. Therefore, gaining insight into the dynamics of SWI is crucial, particularly concerning the various factors that influence the intrusion mechanism. The present study focuses on the experimental simulation of saltwater in freshwater aquifers, considering boundary conditions and density-dependent effects. Two geological scenarios within coastal environments were investigated: First, a uniform, homogeneous case consisting of only sand, and second, a heterogeneous case in which layers of sand, clay, and sand mixed with pebbles are used. During the experiment, DC resistivity sounding data, as part of a widely recognized geophysical method, were collected and subsequently inverted to determine the depth of the freshwater–saltwater interface (FSWI). A finite element analysis was employed to generate numerical models based on experimental feedback. Further, for validation purposes, electrical resistivity tomography (ERT) data were collected from two distinct locations: near the seacoast and an aquaculture area. The ERT results show the presence of salinity intrusion in the study area, attributed mainly to groundwater overpumping and fish farming practices. The experimental findings indicate that the advancement of saltwater is affected by the geological properties of the media they traverse. The porosity (ϕ) and permeability (k) of the geological layer play a crucial role during the passage of saltwater flux into freshwater aquifers. The FSWI deviated along the clay boundary and hindered the easy passage of saltwater into surrounding layers. The alignment of experimental, numerical, and geophysical data suggests that this integrated approach could be valuable for studying SWI and can be applied in different geological settings, including tidal flats and alluvial plains.

HapKled: a haplotype-aware structural variant calling approach for Oxford nanopore sequencing data.

Introduction: Structural Variants (SVs) are a type of variation that can significantly influence phenotypes and cause diseases. Thus, the accurate detection of SVs is a vital part of modern genetic analysis. The advent of long-read sequencing technology ushers in a new era of more accurate and comprehensive SV calling, and many tools have been developed to call SVs using long-read data. Haplotype-tagging is a procedure that can tag haplotype information on reads and can thus potentially improve the SV detection; nevertheless, few methods make use of this information. In this article, we introduce HapKled, a new SV detection tool that can accurately detect SVs from Oxford Nanopore Technologies (ONT) long-read alignment data. Methods: HapKled utilizes haplotype information underlying alignment data by conducting haplotype-tagging using Whatshap on the reads to improve the detection performance, with three unique calling mechanics including altering clustering conditions according to haplotype information of signatures, determination of similar SVs based on haplotype information, and slack filtering conditions based on haplotype quality. Results: In our evaluations, HapKled outperformed state-of-the-art tools and can deliver better SV detection results on both simulated and real sequencing data. The code and experiments of HapKled can be obtained from https://github.com/CoREse/HapKled. Discussion: With the superb SV detection performance that HapKled can deliver, HapKled could be useful in bioinformatics research, clinical diagnosis, and medical research and development.

Discovering Data Domains and Products in Data Meshes Using Semantic Blueprints

Nowadays, one of the greatest challenges in data meshes revolves around detecting and creating data domains and data products for providing the ability to adapt easily and quickly to changing business needs. This requires a disciplined approach to identify, differentiate and prioritize distinct data sources according to their content and diversity. The current paper tackles this highly complicated issue and suggests a standardized approach that integrates the concept of data blueprints with data meshes. In essence, a novel standardization framework is proposed that creates data products using a metadata semantic enrichment mechanism, the latter also offering data domain readiness and alignment. The approach is demonstrated using real-world data produced by multiple sources in a poultry meat production factory. A set of functional attributes is used to qualitatively compare the proposed approach to existing data structures utilized in storage architectures, with quite promising results. Finally, experimentation with different scenarios varying in data product complexity and granularity suggests a successful performance.

Building a searchable online corpus of Australian and New Zealand aligned speech

ABSTRACT Advances in automatic speech recognition technology, increases in bandwidth availability, and the widespread use of video streaming and sharing platforms have opened new horizons for corpus phonetics. CoANZSE Audio, a searchable online version of the Corpus of Australian and New Zealand Spoken English, provides access to over 195 million words of transcribed speech from transcripts of videos uploaded to YouTube by councils and other local government entities in Australia and New Zealand. Audio and forced alignment files are also available, making the resource suitable for the investigation of a range of research questions pertaining to morphosyntax, phonetics, and discourse. The resource, which is freely available via login through CLARIN, Europe’s main language resources infrastructure network, was created through the use of open-source tools and software: yt-dlp, a Python library for collecting data from video and streaming websites; the Montreal Forced Aligner, a recent neural network alignment suite; and Parselmouth-Praat, Python bindings for the Praat acoustic analysis software. The website is powered by BlackLab, which combines a powerful search engine based on Apache Lucene with an intuitive web frontend. CoANZSE Audio may be useful for the investigation of regional differentiation of language features, and with additional annotation, differences in feature use according to social or demographic groups. Recent applications have included studies of double modals, a rare syntactic feature, and apology sequences. The nature of the audio and alignment data may make the resource especially suitable for the study of regional phonetic variation. Furthermore, the methods used to create the resource may be of interest to researchers seeking to adopt a pipeline approach for the creation of specialized corpora from publicly available online content.

Comparison of Perimeter Delineation Methods for Remote Sensing Fire Spot Data in Near/Ultra-Real-Time Applications

Forest fires cause extensive damage to ecosystems, biodiversity, and human property, posing significant challenges for emergency response and resource management. The accurate and timely delineation of forest fire perimeters is crucial for mitigating these impacts. In this study, methods for delineating forest fire perimeters using near-real-time (NRT) remote sensing data are evaluated. Specifically, the performance of various algorithms—buffer, concave, convex, and combination methods—using VIIRS and MODIS datasets is assessed. It was found that increasing concave α values improves the matching percentage with reference areas but also increases the commission error (CE), indicating overestimation. The results demonstrate that combination methods generally achieve higher matching percentages, but also higher CEs. These findings highlight the trade-off between improved perimeter accuracy and the risk of overestimation. The insights gained are significant for optimizing sensor data alignment techniques, thereby enhancing rapid response, resource allocation, and evacuation planning in fire management. This research is the first to employ multiple algorithms in both individual and synergistic approaches with NRT or ultra-real-time (URT) active fire data, providing a critical foundation for future studies aimed at improving the accuracy and timeliness of forest fire perimeter assessments. Such advancements are essential for effective disaster management and mitigation strategies.

Injury-induced Foxm1 expression in the mouse kidney drives epithelial proliferation by a cyclin F-dependent mechanism.

Acute kidney injury (AKI) strongly upregulates the transcription factor Foxm1 in the proximal tubule in vivo, and Foxm1 drives epithelial proliferation in vitro. Here, we report that deletion of Foxm1 either with a nephron-specific Cre driver or by inducible global deletion reduced proximal tubule proliferation after ischemic injury in vivo. Foxm1 deletion led to increased AKI to chronic kidney disease transition, with enhanced fibrosis and ongoing tubule injury 6 weeks after injury. We report ERK mediated FOXM1 induction downstream of the EGFR in primary proximal tubule cells. We defined FOXM1 genomic binding sites by cleavage under targets and release using nuclease (CUT&RUN) and compared the genes located near FOXM1 binding sites with genes downregulated in primary proximal tubule cells after FOXM1 knockdown. The aligned data sets revealed the cell cycle regulator cyclin F (CCNF) as a putative FOXM1 target. We identified 2 cis regulatory elements that bound FOXM1 and regulated CCNF expression, demonstrating that Ccnf is strongly induced after kidney injury and that Foxm1 deletion abrogates Ccnf expression in vivo and in vitro. Knockdown of CCNF also reduced proximal tubule proliferation in vitro. These studies identify an ERK/FOXM1/CCNF signaling pathway that regulates injury-induced proximal tubule cell proliferation.

Optimal transport for automatic alignment of untargeted metabolomic data.

Untargeted metabolomic profiling through liquid chromatography-mass spectrometry (LC-MS) measures a vast array of metabolites within biospecimens, advancing drug development, disease diagnosis, and risk prediction. However, the low throughput of LC-MS poses a major challenge for biomarker discovery, annotation, and experimental comparison, necessitating the merging of multiple datasets. Current data pooling methods encounter practical limitations due to their vulnerability to data variations and hyperparameter dependence. Here, we introduce GromovMatcher, a flexible and user-friendly algorithm that automatically combines LC-MS datasets using optimal transport. By capitalizing on feature intensity correlation structures, GromovMatcher delivers superior alignment accuracy and robustness compared to existing approaches. This algorithm scales to thousands of features requiring minimal hyperparameter tuning. Manually curated datasets for validating alignment algorithms are limited in the field of untargeted metabolomics, and hence we develop a dataset split procedure to generate pairs of validation datasets to test the alignments produced by GromovMatcher and other methods. Applying our method to experimental patient studies of liver and pancreatic cancer, we discover shared metabolic features related to patient alcohol intake, demonstrating how GromovMatcher facilitates the search for biomarkers associated with lifestyle risk factors linked to several cancer types.

Optimal transport for automatic alignment of untargeted metabolomic data

Untargeted metabolomic profiling through liquid chromatography-mass spectrometry (LC-MS) measures a vast array of metabolites within biospecimens, advancing drug development, disease diagnosis, and risk prediction. However, the low throughput of LC-MS poses a major challenge for biomarker discovery, annotation, and experimental comparison, necessitating the merging of multiple datasets. Current data pooling methods encounter practical limitations due to their vulnerability to data variations and hyperparameter dependence. Here, we introduce GromovMatcher, a flexible and user-friendly algorithm that automatically combines LC-MS datasets using optimal transport. By capitalizing on feature intensity correlation structures, GromovMatcher delivers superior alignment accuracy and robustness compared to existing approaches. This algorithm scales to thousands of features requiring minimal hyperparameter tuning. Manually curated datasets for validating alignment algorithms are limited in the field of untargeted metabolomics, and hence we develop a dataset split procedure to generate pairs of validation datasets to test the alignments produced by GromovMatcher and other methods. Applying our method to experimental patient studies of liver and pancreatic cancer, we discover shared metabolic features related to patient alcohol intake, demonstrating how GromovMatcher facilitates the search for biomarkers associated with lifestyle risk factors linked to several cancer types.