01
Data Lab — Exploration & QC
Get an instant overview of your dataset: sample count, feature distribution, missing value patterns visualised as an interactive heatmap, coefficient of variation per feature, and class balance. An outlier detection module using Isolation Forest flags suspicious samples before any downstream analysis.
◆ Missing value heatmap & per-sample % report
◆ Distribution plots & CV analysis
◆ Isolation Forest outlier detection
◆ Class balance visualisation (SMOTE / ADASYN ready)
◆ Sample rename, edit & metadata (_meta columns)
Missing heatmapIsolation ForestCV analysisClass balance
02
Preprocessing Pipeline
A complete, ordered pipeline: filter by missing-value threshold, impute with KNN or median, normalize (log₂, Z-score, quantile, robust), correct batch effects with ComBat, apply variance filtering, and validate the result in a post-QC dashboard — all in a single step with data-driven auto-suggestions.
◆ KNN imputation · median · min/2
◆ Log₂ · Z-score · quantile · robust normalization
◆ ComBat batch correction (NeuroCombat)
◆ Variance threshold filtering
◆ Post-QC validation dashboard
◆ SMOTE · ADASYN class balancing
ComBatKNN imputationLog₂ · Z-scoreSMOTE · ADASYN
03
Data Visualisation
Every plot is fully interactive (Plotly) — zoom, pan, hover tooltips, lasso selection, publication-ready export at 2× resolution. Dimensionality reduction, clustering heatmaps, signal profiles and pairwise comparisons all in one module.
◆ PCA · UMAP · t-SNE with class overlays
◆ Hierarchical clustering heatmap
◆ Correlation matrix & cosine similarity heatmap
◆ Feature distribution: violin, boxplot, density
◆ Signal profile (multi-feature line plot)
◆ PNG · SVG export at publication resolution
PCA · UMAP · t-SNEHeatmap clusteringCorrelationInteractive Plotly
04
Differential Analysis & Biomarker Discovery
Multi-group and pairwise statistical comparisons with FDR-corrected p-values. The volcano plot supports multi-class mode, customisable fold-change and p-value thresholds, and direct export of significant features into enrichment. UpSet & Venn plots reveal exclusive and shared features across conditions.
◆ Volcano plot — binary & multi-class, interactive
◆ t-test · Wilcoxon · Mann-Whitney · ANOVA · Kruskal-Wallis
◆ FDR correction (BH, Bonferroni)
◆ Heatmap of significant features with clustering
◆ Venn & UpSet plots — exclusive/shared feature sets
◆ Feature boxplots per group with significance bars
Volcano plotANOVA · WilcoxonFDR correctionVenn · UpSet
05
AI Modeling — Classification & Regression v1.2
Train and evaluate machine learning and deep learning models for classification or continuous regression (new in v1.2). Full pipeline: cross-validation, hyperparameter tuning, confusion matrices, ROC curves, R² and RMSE metrics — plus explainability via SHAP & LIME at sample and global level. Models are exportable as .pkl for real-time inference.
◆ Random Forest · XGBoost · SVM · Gradient Boosting · k-NN
◆ Logistic & Linear Regression · PLS-DA
◆ MLP (Deep Learning) — configurable layers, dropout, early stopping
◆ Cross-validation · hyperparameter tuning · model comparison chart
◆ SHAP beeswarm · bar · force plots
◆ LIME explanations (sample-level & global)
◆ Regression: R² · RMSE · residual plots new
◆ Real-time prediction on new unseen samples
XGBoost · RF · SVMMLP deep learningSHAP & LIMERegression ✦new.pkl export
06
ORA + GSEA Pathway Enrichment v1.2
GSEA is new in v1.2, joining the existing Over-Representation Analysis. Both methods run across 100+ databases via gseapy. The smart auto-detection engine pre-fills the feature input from any upstream result: click a heatmap cluster, a volcano significant set, a Venn exclusive zone — or paste manually.
◆ GSEA — ranked gene lists from any Profiler output new
◆ ORA — over-representation across 100+ databases
◆ GO Biological Process · Molecular Function · Cellular Component
◆ Reactome · KEGG · WikiPathways · DrugBank · MSigDB
◆ Enrichment bar · dot plot · heatmap · gene–pathway network
◆ Auto-import from volcano · heatmap · Venn · UpSet
GSEA ✦newORAGO · Reactome · KEGG100+ databasesAuto-import
07
Survival Analysis
Model time-to-event outcomes from omics features. Kaplan–Meier estimator with log-rank test, Cox proportional hazards regression with forest plot, and risk group stratification. All curves are interactive and publication-ready.
◆ Kaplan–Meier estimator per group
◆ Log-rank test with p-value annotation
◆ Cox model — forest plot of hazard ratios
◆ Risk stratification from continuous features
◆ Survival table & at-risk annotations
Kaplan–MeierCox modelForest plotRisk stratification
08
Longitudinal Analysis v1.2
A dedicated module for repeated-measures and time-series omics data. Load a dataset with Subject_ID and Time columns and instantly visualise molecular trajectories, compare group dynamics, and run time-point statistics.
◆ Trajectory plots per feature & per subject
◆ Group-level dynamics with confidence intervals
◆ Repeated-measures ANOVA / mixed models
◆ Time-point pairwise comparisons
◆ Compatible with all omics types + lipidomics time-series
Trajectory plotsMixed modelsRepeated measuresSubject_ID + Time
09
HTML Report Generation v1.2
At any point in your session, generate a complete, self-contained HTML report capturing every analysis: all interactive Plotly figures, statistical tables, model metrics, enrichment results and a table of contents. The file opens in any browser with no dependencies — shareable, archivable, publication-ready.
◆ All session plots embedded as interactive Plotly HTML
◆ Statistical tables, model metrics, enrichment results
◆ Auto-generated table of contents
◆ Self-contained — works offline, no server needed
◆ Timestamped & branded with Profiler + PRISM
One-click exportSelf-contained HTMLAll plots includedOffline-ready
10
Desktop & MSI2Profiler
Run the complete Profiler platform locally — no internet, no account, no upload size limit. MSI2Profiler is a companion desktop tool for preprocessing Mass Spectrometry Imaging data (.imzML), defining ROIs and exporting matrices directly importable into Profiler.
◆ Windows 10/11 · macOS 10.15+ · Linux Ubuntu 20.04+
◆ Full feature parity with the web version
◆ No upload size limit · runs on local compute
◆ MSI2Profiler — imzML · MALDI-MSI · DESI-MSI ·SpiderMass-MSI
◆ ROI definition & concatenation → Profiler matrix
Windows · macOS · LinuxMSI2ProfilerimzML · MALDINo limits