v1.2
Regression modeling in supervised learning · GSEA added alongside ORA with smart feature auto-detection from volcano, heatmap, Venn/UpSet · Longitudinal analysis module · All plots fully interactive · Instant HTML reports · Improved UI · TSV tab-delimiter fix · Extended format support · MSI2Profiler desktop tool.
Data input & preprocessing
01
Multi-Omics Data Support v1.2
Proteomics, metabolomics, lipidomics, transcriptomics, genomics. Auto-detection extended to protein group files — MaxQuant
proteinGroups.txt, DIA-NN pg_matrix.tsv, Spectronaut, FragPipe, Perseus, DESeq2/edgeR, Salmon, kallisto, MetaboAnalyst, XCMS, MZmine.MaxQuantDIA-NNSpectronautDESeq2Auto-detect
02
Preprocessing Pipeline v1.2
Redesigned with smart auto-suggestions per dataset type. Normalize, filter, impute, correct batches with ComBat, handle zero-inflation — with a full post-QC dashboard. Clean & Edit changes now propagate correctly to all derived datasets.
ComBatSMOTEKNN imputationSmart suggestions
03
Interactive Visualization v1.2
All figures in v1.2 are fully interactive — zoom, pan, hover tooltips, lasso selection. PCA, UMAP, t-SNE, heatmaps, volcano plots, violin, boxplots, correlation matrices. Export PNG, SVG, PDF at 2× resolution.
PCA · UMAP · t-SNEHeatmapsVolcanoAll interactive
04
HTML Report Generation New
Generate a complete, self-contained HTML report at any point in your analysis. All preprocessing summaries, QC plots, statistical comparisons, ML results, enrichment tables — in a single offline-readable file shareable with collaborators.
Instant exportSelf-containedOffline-readable
AI & Statistical Modeling
05
Machine Learning Regression New
Classification and regression (new in v1.2) with Random Forest, XGBoost, SVM, Gradient Boosting, k-NN. Cross-validation, hyperparameter tuning, confusion matrices, R² and RMSE. Save and reload models (.pkl).
RegressionXGBoostRandom ForestCross-val
06
Deep Learning v1.2
MLP networks for classification and regression. Configure layers, units, dropout. Monitor training curves in real-time. Early stopping. Export models for external prediction pipelines.
RegressionMLPEarly stoppingExport .pkl
07
Biomarker Discovery v1.2
Identify discriminant features via volcano plots, fold-change, ANOVA/t-test. Improved SHAP visualisations and LIME explanations at global and per-sample level. Selected features feed directly into the enrichment module.
SHAP (improved)LIMEVolcano → Enrichment
08
Group Comparisons
Pairwise and multi-group statistical tests — t-test, ANOVA, Wilcoxon, Mann-Whitney with FDR correction. Venn and UpSet plots for exclusive/shared feature discovery, with direct auto-feed into the enrichment module.
Venn / UpSet → EnrichmentBH-FDRBonferroni
Pathway Enrichment
ORA + GSEA — 100+ Databases GSEA New
v1.2 adds Gene Set Enrichment Analysis (GSEA) alongside the existing Over-Representation Analysis (ORA). Both methods across 100+ databases including GO, Reactome, KEGG, DrugBank, WikiPathways, MSigDB and Disease Ontology.
Features are auto-detected from the source — select a cluster in a heatmap, significant hits from a volcano plot, an exclusive zone in a Venn/UpSet plot, or paste a manual list. Profiler pre-fills the enrichment input automatically.
Volcano Plot
→
Heatmap Cluster
→
Venn / UpSet
→
Manual List
→
ORA + GSEA
GSEAGO BP/MF/CCReactomeKEGGDrugBankMSigDBAuto-detect features
Bioinformatics & Clinical
09
Survival Analysis
Kaplan–Meier curves and Cox proportional hazards models. Stratify patients by molecular signatures. Log-rank test for group comparison. Integrates with
survival_meta and pfs_meta columns.Kaplan–MeierCox regressionLog-rank
10
Longitudinal Analysis New
Dedicated module for repeated-measures omics data. Track molecular trajectories across time points, compare group dynamics, visualise individual subject profiles. Requires
Subject_ID and Time columns.New in v1.2Trajectory plotsRepeated measures
11
MSI2Profiler New Tool
Desktop tool for preprocessing Mass Spectrometry Imaging .imzML files. TIC/Median/RMS normalization, spectral binning, ROI concatenation. Export directly to Profiler-ready CSV.
imzML · MALDI · DESIWindows app
12
Data Privacy & Desktop
No server-side storage — all data processed in-session and auto-deleted on logout. Desktop version runs entirely offline: no internet, no upload limits, 100% local processing. Windows, macOS, Linux.
No data storedOffline desktopWin · macOS · Linux
Published in Bioinformatics (Oxford, 2026). Please cite the paper if you use Profiler. doi:10.1093/bioinformatics/btaf644