Skip to content

About

an interactive proteomics analysis tool designed to streamline cell-surface protein discovery from mass spectrometry datasets.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

surfaceome_explorer

The plasma membrane serves as the interface between the cell and its microenvironment, directing processes such as cell-cell communication, immune recognition, adhesion, and signal transduction.

Extracellular cell surface proteins are highly accessible to therapeutic antibodies and engineered immune cells; they represent an important source of biomarkers and immunotherapy targets. Yet, distinguishing proteins localized on the extracellular cell surface from those that are membrane-associated or more broadly annotated as plasma membrane proteins often requires manually integrating multiple annotation resources and statistical analyses.

No single reference resource completely captures the complexity of plasma membrane protein localization. Surfaceome Explorer integrates these complementary resources into a unified analytical workflow to prioritize candidate proteins for downstream experimental validation using methods such as flow cytometry, immunofluorescence microscopy, proximity labeling, or targeted mass spectrometry.

Note: This is an exploratory data analysis tool for hypothesis generation. Protein annotations are based on curated public databases and should be interpreted alongside experimental evidence.

landing page

Landing Page


overview

Surfaceome Explorer is a web-based application for interpreting quantitative proteomics datasets by integrating surfaceome annotation, differential abundance analysis, visualization, and candidate prioritization.

The tool accepts reports from secondary mass spectrometry analysis (peptide reports from DIA-NN, FragPipe, MaxQuant, etc.) and cross-references detected proteins against three curated surfaceome reference databases. It performs group-wise differential abundance analysis using Welch's t-test with a Benjamini–Hochberg false discovery rate correction, and presents results through an interactive dashboard that requires no installation, server, or programming.

The entire application runs locally in the user's browser. No data is uploaded to any external server.

data analysis demo

demo


key features

  • Zero-installation deployment — a single .html file that runs in any modern browser without dependencies, server access, or internet connection after initial load
  • Multi-database surfaceome annotation — cross-references detected proteins against SURFY (in silico surfaceome), UniProt Cell Membrane (KW-1003), and the Membrane-Associated dataset of Almén et al.
  • Proportional Venn diagrams — area-proportional SVG diagrams showing database co-annotation overlap and per-category group detection overlap, with hover interaction and click-to-filter
  • Differential abundance analysis — pairwise Welch's t-test between experimental groups, with Benjamini–Hochberg FDR correction; results presented in sortable tables and a volcano plot
  • Volcano plot — interactive scatter plot with configurable category filters, persistent gene-name labels on the most significant proteins, and PNG export
  • Priority scoring — the tool generates a composite score integrating database overlap, detection frequency, and statistical significance to rank candidate surface proteins
  • Condition-specific protein detection — identifies proteins detected exclusively in one experimental group
  • All-in-one export — per-table CSV downloads, PNG exports for all figures, and a formatted PDF summary report
  • No data leaves the browser — all computation is performed client-side in JavaScript; raw data is never transmitted

workflow

Quantitative proteomics input
        │
        ▼
  ┌─────────────────────────────┐
  │   Upload & Parse (Tab 1)    │  ← Upload processed proteomics file
  │   Sample / group mapping    │
  └──────────────┬──────────────┘
                 │
                 ▼
  ┌─────────────────────────────┐
  │  Database Annotation        │  ← Cross-reference against SURFY,
  │  (SURFY · UniProt · MA)     │    UniProt KW-1003, Almén MA dataset
  └──────────────┬──────────────┘
                 │
                 ▼
  ┌─────────────────────────────┐
  │  Differential Analysis      │  ← Welch's t-test, BH-FDR correction
  │  (pairwise, UniProt PM set) │    per UniProt Cell Membrane protein
  └──────────────┬──────────────┘
                 │
        ┌────────┴────────┐
        ▼                 ▼
  Venn diagrams      Volcano plot
  Ranked tables      Priority scores
  PDF report         CSV exports

installation

No installation is required. Surfaceome Explorer is a single HTML file that runs entirely in the browser.

to use:

  1. Download surfaceome-explorer-v2.html from the Releases page (or clone this repository).
  2. Open the file in a web browser (Chrome recommended).
  3. Upload your data file using the interface on the first tab.

No Python, R, Node.js, or any other runtime is required.

browser compatibility: The application has been developed and tested primarily in Google Chrome. Functionality in other browsers is expected but has not been exhaustively verified across all features.


required input format

Surfaceome Explorer currently accepts tab-separated value (.tsv) or comma-separated value (.csv) files produced by standard label-free quantification (LFQ) pipelines. The expected format is a protein-level quantification matrix in which:

  • Each row represents a unique protein (identified by UniProt accession)
  • Each column (beyond the identifier columns) represents a sample or replicate
  • Abundance values should be log2-transformed LFQ intensities or equivalent pre-processed values; missing values should be represented as empty cells or NaN

minimum required columns

Column Description
Protein IDs or Accession UniProt accession (e.g., P12345)
One or more sample columns Numeric abundance values per replicate

group / sample mapping

After upload, a mapping interface lets you assign each sample column to an experimental group (e.g., Control, Treatment). At least two groups, each with at least two replicates, are required for differential analysis.

compatibility note: This input format is consistent with output from MaxQuant (proteinGroups.txt, filtered), Perseus, and similar tools. Direct parsing of raw mass spectrometry files (mzML, raw, wiff) is not currently supported. Future versions are planned to include direct mzML parsing; early-stage mzML reading has been prototyped and tested on small example files, but broader compatibility has not yet been validated.


analysis pipeline

1. protein filtering and annotation

After upload, detected proteins are filtered to remove known contaminants and reverse-sequence hits (columns flagged in MaxQuant output). Each remaining protein is cross-referenced against the three surfaceome reference databases using UniProt accession matching. Proteins are assigned to one or more database categories:

  • SURFY — computationally predicted cell surface proteins
  • UniProt Cell Membrane (KW-1003) — experimentally annotated plasma membrane proteins from UniProt
  • Membrane-Associated (MA) — a curated membrane-associated dataset

2. group-wise abundance summarization

For each protein and experimental group, mean abundance and standard deviation are calculated from the available replicates. Proteins detected in fewer than 50% of replicates within a group are flagged as low-confidence detections in that group.

3. differential abundance analysis

Pairwise differential analysis is performed for each group comparison across the set of UniProt Cell Membrane (KW-1003) proteins. For each protein detected in both groups:

  • fold change is computed as log₂(mean group B / mean group A)
  • statistical significance is assessed using Welch's t-test (unequal variance), appropriate for small and unequal sample sizes
  • multiple testing correction is applied using the Benjamini–Hochberg procedure to compute per-protein FDR values

Proteins with p < 0.05 are highlighted in the differential tables. Volcano plot coloring uses p < 0.05 as the significance threshold.

4. priority scoring

A composite priority score is assigned to each detected surface protein, integrating:

Component Points
SURFY membership +3
UniProt Cell Membrane membership +2
Membrane-Associated membership +1
Differential abundance (p < 0.05, |log₂FC| > 1) +3
High mean abundance (top quartile) +2
Detected in all replicates +2

This score is intended to help prioritize candidates for follow-up experiments and does not represent a validated biological ranking.


reference databases

Database Description Version / Source
SURFY In silico predicted human surfaceome based on machine-learning classification of signal peptides, transmembrane domains, and GPI anchors Bausch-Fluck et al. (2018)
UniProt Cell Membrane (KW-1003) Proteins manually annotated with the UniProt keyword Cell membrane UniProt release used at time of database build; see databases/ folder
Membrane-Associated (MA) A curated set of membrane-associated human proteins Almén et al. (2009)

Database annotation files are bundled within the application at build time. Future releases will provide a mechanism for users to supply updated database files.


downloadable outputs

All exports are generated client-side and downloaded directly to the user's machine.

Output Format Description
All detected surfaceome proteins .csv Full protein table with abundance, database membership, and priority score
Upregulated proteins .csv Proteins with log₂FC > 0, sorted by fold change
Downregulated proteins .csv Proteins with log₂FC < 0, sorted by fold change
Volcano plot .png High-resolution PNG of the current volcano plot view
Category Venn diagrams .png PNG export of each group-overlap Venn diagram
Cytoscape node table .csv Protein table formatted for direct import into Cytoscape
Summary report .pdf Formatted PDF summarizing metadata, QC metrics, key findings, and all figures

biological interpretation

Surfaceome Explorer is designed to assist in the exploratory analysis of cell-surface proteomics data. Key outputs to consider:

Database overlap measures the degree of co-annotation across databases for your detected proteins. A large proportion of proteins annotated by SURFY but absent from UniProt Cell Membrane may indicate predicted but not yet experimentally confirmed surface proteins. These represent biologically interesting candidates but should be interpreted cautiously.

High group-overlap Venn diagrams (>85% shared proteins) indicate that the experimental conditions share a largely conserved surface proteome. Differences in the small unique-to-group fractions may be biologically meaningful even when modest in number.

Volcano plot significance uses an uncorrected p-value threshold of 0.05 for visualization purposes. For biological conclusions, we recommend cross-referencing with the FDR-adjusted values in the differential tables. Treat proteins near the significance boundary as preliminary leads.

Priority scores are heuristic and unvalidated. They are intended to guide experimental prioritization, not to replace biological judgment or rigorous statistical inference.

This tool was developed with Claude Code; the tool design, analysis workflows, and statistical logic are my own.

references

Reference databases used in Surfaceome Explorer are derived from the following published resources:

  • SURFY: Bausch-Fluck D, Goldmann U, Trefzer A, et al. (2018). The in silico human surfaceome. PNAS 115(46):E10988–E10997. https://doi.org/10.1073/pnas.1808790115
  • UniProt: The UniProt Consortium (2023). UniProt: the Universal Protein Knowledgebase in 2023. Nucleic Acids Research 51(D1):D523–D531. https://doi.org/10.1093/nar/gkac1052
  • Membrane-Associated dataset: Almén MS, Nordström KJ, Fredriksson R, Schiöth HB (2009). Mapping the human membrane proteome: most human membrane proteins can be classified by function and evolutionary origin. BMC Biology 7:50. https://doi.org/10.1186/1741-7007-7-50

statistical methodology:

  • Welch's t-test: Welch BL (1947). The generalization of "Student's" problem when several different population variances are involved. Biometrika 34(1–2):28–35.
  • Benjamini–Hochberg FDR: Benjamini Y, Hochberg Y (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Royal Statistical Society B 57(1):289–300.

About

an interactive proteomics analysis tool designed to streamline cell-surface protein discovery from mass spectrometry datasets.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages