FlowJEM

FlowJEM
Name	FlowJEM
Developer	Independent Laboratory
Released	2024
Latest release	1.2
Programming language	C++, Python
Operating system	Linux, macOS
License	Proprietary

FlowJEM

FlowJEM is a software platform for high-throughput flow cytometry data management and joint embedding visualization. The system integrates data ingestion, computational modeling, and interactive visualization to support laboratory workflows in translational research, clinical trials, and large consortium projects. FlowJEM emphasizes reproducibility, scalability, and interoperability with established bioinformatics ecosystems.

Overview

FlowJEM combines components for data parsing, single-cell feature extraction, dimensionality reduction, batch correction, and visualization. It interoperates with formats and tools originating from institutions such as Broad Institute, European Bioinformatics Institute, National Institutes of Health, Wellcome Trust Sanger Institute, and Stanford University. The platform supports pipelines built around pipelines from projects like Human Cell Atlas and integrates with repositories such as ArrayExpress and Gene Expression Omnibus for metadata linkage. FlowJEM is positioned alongside software created by groups at Massachusetts Institute of Technology, Harvard University, University of California, San Francisco, Johns Hopkins University, and Cold Spring Harbor Laboratory.

History and Development

FlowJEM's development began as a collaboration between a translational research group and computational teams influenced by methods from Broad Institute labs and analytic frameworks documented by European Bioinformatics Institute workshops. Early prototypes were tested in consortium studies coordinated with Human Cell Atlas contributors and pilot cohorts enrolled through networks like Clinical and Translational Science Awards funded by National Center for Advancing Translational Sciences. Subsequent releases incorporated algorithms inspired by work cited in publications from Stanford University, Massachusetts Institute of Technology, Harvard Medical School, University of Cambridge, and University of Oxford. Commercialization and support models drew on practices from companies associated with Genentech, Roche, Illumina, 10x Genomics, and BD Biosciences.

Architecture and Components

FlowJEM is structured as modular services and native libraries. Core components parallel architectures developed at Google and Microsoft Research for scalable compute orchestration, and borrow serialization patterns used in projects at Facebook AI Research and DeepMind. Major modules include a data loader compatible with standards endorsed by Mass Cytometry Standards Consortium and data models referencing schemas from BioSamples and FAIRsharing. Computational engines implement dimensionality reduction methods originally popularized in publications from University of Toronto and Princeton University groups, while visualization components draw UX design cues from implementations in Tableau and Plotly libraries used by teams at Allen Institute for Brain Science.

Operation and Workflow

Operationally, FlowJEM ingests cytometry files from instruments produced by manufacturers such as BD Biosciences, Beckman Coulter, Sony Biotechnology, Thermo Fisher Scientific, and Cytek Biosciences. Users supply sample metadata consistent with submission protocols used by National Cancer Institute studies and clinical trial registries like ClinicalTrials.gov. The workflow executes preprocessing steps akin to those described by labs at Yale University and University of Pennsylvania: compensation, gating, transformation, and quality control, then performs embedding using algorithms that trace lineage to methods from Johns Hopkins University and Columbia University. Result artifacts are compatible with downstream analysis in environments developed at RStudio, Bioconductor, and Python Software Foundation ecosystems.

Applications and Use Cases

FlowJEM is applied in translational immunology studies led by groups at Imperial College London and Weill Cornell Medicine, in cancer immunotherapy trials coordinated with Memorial Sloan Kettering Cancer Center and Dana-Farber Cancer Institute, and in vaccine response studies associated with Bill & Melinda Gates Foundation collaborations. It is used for cell population discovery in multicenter consortia such as Human Cell Atlas and disease-focused initiatives funded by Wellcome Trust and European Commission programs. Public health agencies including Centers for Disease Control and Prevention and World Health Organization have evaluated comparable platforms for outbreak response and surveillance pipelines employed in collaborations with Johns Hopkins Bloomberg School of Public Health.

Performance and Evaluation

Performance benchmarks for FlowJEM were conducted on compute clusters similar to architectures published by Lawrence Berkeley National Laboratory and Argonne National Laboratory. Evaluation protocols compared embedding stability, batch correction effectiveness, and throughput against reference implementations developed at Broad Institute, EMBL-EBI, and University of California, San Diego. Metrics reported in independent assessments reflect trade-offs studied in papers from Massachusetts Institute of Technology, University of Washington, and University of California, Berkeley groups. Validation exercises were performed using public datasets curated by European Bioinformatics Institute and consortium datasets shared by Human Cell Atlas partners.

Limitations and Challenges

FlowJEM faces limitations common to high-dimensional single-cell platforms encountered in studies from Stanford University and Harvard Medical School: sensitivity to staining variability from manufacturers such as BD Biosciences and Beckman Coulter, computational costs comparable to cloud services offered by Amazon Web Services, Google Cloud Platform, and Microsoft Azure, and regulatory considerations overseen by agencies like Food and Drug Administration and European Medicines Agency. Ongoing challenges include harmonizing heterogeneous metadata standards advocated by Global Alliance for Genomics and Health and scaling visual analytics workflows to cohort sizes examined by projects at Wellcome Sanger Institute and Human Cell Atlas.

Category:Bioinformatics software