> For the complete documentation index, see [llms.txt](https://docs.owkin.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.owkin.com/terms-and-support/support/changelog.md). # Changelog ### July 2026 ### **2026-07-23** **Copy-number variation analysis (InferCNV)** You can now infer large-scale copy-number gains and losses from single-cell or spatial transcriptomics data against a normal reference. Results display as a chromosome heatmap, with optional clustering of the inferred profiles into putative tumour clones. **Oncoprint visualisation** K-Pro can now generate an oncoprint showing the mutation and copy-number-alteration landscape across a cohort for a chosen gene panel, with samples ordered so co-altered cases group together. **Classification result visualisations** Classification analyses now include built-in visualisations of their results. **Gantt chart export improvements** Resolved an issue where exported Gantt charts could lose bar labels, and added an overall-survival / progression-free-survival maturity legend to PowerPoint exports. **Smoother response rendering** Improved markdown formatting and spacing while responses stream, and fixed a case where the fullscreen panel stayed open after being closed. ### 2026-07-08 **Dataset access is now project-scoped** Reading a dataset is now limited to members of that dataset's project, giving clearer, project-based control over who can see what. **Clearer plots with legends and wide labels** Charts with legends or wide axis labels now keep their full margins instead of clipping, fixing cases where swimlane legends were missing from Gantt charts. ### 2026-07-02 **Pathway and gene-signature enrichment analysis** K-Pro can now score pathways and gene signatures across your cohort. Single-sample scoring (ssGSEA) returns an enrichment score per sample for each signature, comparable across patient groups, and pre-ranked GSEA scores a ranked gene list against gene sets, returning a normalized enrichment score and significance per pathway. Signatures can be supplied manually or taken from curated pathway databases. ### June 2026 ### 2026-06-26 **Interactive whole-slide viewer** Pathology Explorer now lets you open a whole-slide image and pan, zoom, and view it fullscreen. Tiles stream on the fly, so large slides open in seconds instead of waiting on a full download. **Streaming responses** Answers and result cards now appear progressively as they are generated, so you can start reading and reviewing results while the rest is still being produced. **Richer data catalog** The Data Generation catalog is now organized around curated profiles, product details no longer repeat across panels, and the map auto-zooms to your current filters as you refine them. ### 2026-06-19 **scRNAseq Pseudobulk Differential Expression Analysis** K-Pro can now compare gene expression between two patient groups across all cell types aggregated per patient, or within a specific cell type such as malignant cells.. **Export for reports and plots** Reports and plots now include an export button directly in the interface. ### 2026-06-12 **Pathology Explorer** Explore whole-slide pathology images directly in K-Pro, view slide thumbnails, inspect image tiles alongside model predictions, and overlay spatial maps.\ \ **Biomarker Extraction** A new analysis identifies a small set of features that best separates a chosen group of patients and presents them as a simple decision tree, giving you candidate biomarkers you can act on.\ \ **Analysis Rationale** Every plot now includes a short explanation of why that analysis answers your question.\ \ Resolved an issue where logout could be blocked during onboarding. ### May 2026 ### 2026-05-28 **Unified Agent Orchestration** All K-Pro users now run on the next-generation orchestration pipeline, producing more direct and consistent answers across multi-step questions. **Gene–Pathway Lookup** K-Pro can now map a gene to the biological pathways it belongs to, and list the genes within a given pathway, using curated MSigDB gene sets. **Consensus Clustering for Patient Grouping** Multi-omics analyses now include consensus agglomerative clustering to group patients, returning a consensus matrix and cluster labels. Silhouette and purity scores are reported alongside scatter and consensus clustering analyses to help assess cluster quality. **Gene Availability Checks** Analyses now verify that each requested gene is present in the dataset and notify you when one is missing, instead of failing silently. **Pathology Explorer: Larger Slide Uploads** Whole-slide images up to 4 GB are now supported (previously 2 GB), preventing large slides from being silently rejected. **Clearer Context Compaction** When long conversations are summarised to stay within context limits, a live progress bar is now shown instead of the raw summariser output. ### 2026-05-20 **Pathology Explorer** Flexible cohort name formats Pathology Explorer tools now accept both TCGA-LUAD and TCGA\_LUAD cohort name formats, preventing silent tool call failures when the standard hyphenated format is used. **Multi-omics** Corrected oncogenic variant handling Mutation frequency analysis now correctly treats the oncogenic column as a boolean flag, fixing plots that previously misgrouped oncogenic vs. non-oncogenic variants. **Login reliability** Resolved a rare login loop that could occur when multiple authentication requests fired in parallel. #### 2026-05-06 **Prior Knowledge Database: Tool Restored** Fixed an outage affecting the Bioknowledge Navigator / Prior Knowledge Database tool that prevented queries against curated knowledge sources between late April and early May. Authentication to underlying data has been hardened to prevent recurrence. **Claude 4.x Model Upgrade** All K-Pro services (Bioknowledge Navigator, Multi-Omics, orchestration) now run exclusively on Claude 4.x. Earlier Sonnet 3 / 3.5 / 3.7 and Haiku 3 references have been deprecated, providing more capable reasoning across the platform. ### April 2026 #### 2026-04-28 **Consensus Replaces Literature Review** A new evidence search tool, Consensus, is now live for K-Pro Free users. Compared to the previous Literature Review service: significantly wider research article coverage, with most articles accessible as full text rather than abstracts only; new filters for recency (e.g. publications from the last year) and top-tier journals. The Explore Data tab now points to Consensus as the literature source. #### 2026-04-27 **Improved Agent Reasoning for K-Pro Free** K-Pro Free users now run on a next-generation agentic pipeline with flat orchestration. The new pipeline streamlines how queries are routed across tools, producing more direct, consistent answers — especially for multi-step clinical questions. Plotter generation has been updated in parallel: cleaner clinical summaries, better population plots, and improved alignment between the data you ask about and the visualization returned. **Onboarding Refresh** The post-login onboarding experience has been redesigned with a single-video modal and a new onboarding quiz that captures user persona for tailored prompt suggestions. Both new and returning users see the redesign. **Personalized Suggested Prompts** Suggested prompts on the chat landing page are now personalized based on user persona, surfacing more relevant starter queries. **Longer Conversations Stay Coherent** New context compaction automatically summarises older conversation history when approaching the token limit, preserving context across long analytical sessions. **Pathology Explorer: GigaTIME Immune Biomarker Features** New tools expose GigaTIME virtual multiplex immunofluorescence (mIF) data for TCGA-LUAD: 21 mean channel expression values per slide via `get_gigatime_features`, marker grid display via `display_gigatime_map`, and HE-vs-marker side-by-side overlays via `display_gigatime_overlay`. Survival analysis and cohort description tools have been extended to support categorical Kaplan-Meier stratification on Gigatime data. #### 2026-04-09 **Streaming Responses** Agent responses now stream in real time as they are generated, reducing perceived latency for long answers. **Explore Data: More Datasets and Refined Filters** Additional datasets are now surfaced in the Explore Data tab, with refined display and filtering for inclusion/exclusion (I/E) criteria. **MOSAIC Data Access Clarity** K-Pro now correctly explains MOSAIC data access: only MOSAIC Window (a 60-patient public subset) is hosted on EGA. Access to the full MOSAIC dataset requires a data licensing conversation with Owkin. #### 2026-04-03 **TA-Aware Population Plots** Clinical summaries now adapt to the dataset's Therapeutic Area. Oncology queries surface a clinical summary table, an indication-specific summary, and Kaplan-Meier curves for Overall Survival (OS) and Progression-Free Survival (PFS). Immunology & Inflammation studies (Crohn's disease, UC) display a clinical summary table alongside a sample-level disease activity score summary and biopsy site information.\ \ **Authentication Upgrade:** Login flows have been migrated from Amplify to a standard OIDC protocol, improving security and reliability. Microsoft Entra ID is now supported as an identity verifier for enterprise environments.\ \ **Chat History: Infinite Scrolling** The chat history page now supports infinite scrolling for easier navigation through long conversation histories. ### March 2026 #### 2026-03-20 **Therapeutic-Area-Aware Clinical Summaries** Clinical summaries now automatically select the most relevant plots based on the therapeutic area — Kaplan-Meier curves for oncology, variable descriptions for immunology and inflammation datasets. **Variant-Level Table** A new variant-level table is now available in multi-omics analyses, enabling exploration of genomic variant data at the individual variant level. **Procedure Event Table** Patient procedure events are now available as a dedicated table, expanding the clinical data accessible for analysis. #### 2026-03-02 **K Data Catalog: Improved Dataset Discovery** Enhanced dataset metadata and improved navigation make it easier to explore and understand available datasets directly from the platform. **Multi-Omics: Cross-Center Cohort Grouping** Multi-center cohorts can now be grouped at the indication level, removing center-level fragmentation for cleaner population-level analyses. *** ### February 2026 #### 2026-02-18 **Clinical Summaries: Multi-Dataset Queries Fixed** Resolved an issue preventing clinical summaries from querying across multiple datasets in a single session. #### 2026-02-12 **Improved PDF Report Export** Reports now render special characters correctly and plots are placed in order for more accurate exports. *** ### January 2026 #### 2026-01-15 **Pathology MCP Explorer: Tile-Level Visualization** * Display tiles with overlay of cell predictions. * 13 MCP tools total (up from 12) #### 2026-01-14 **Improved chart formatting** Chart titles and axis labels now use consistent capitalization across all visualizations. ### December 2025 #### 2025-12-31 **Pathology MCP Explorer: Tile-Level Analysis** * Filter and sort individual tiles within a slide by any histomics feature. * Get detailed statistics (mean, std, min, max, percentiles) for any feature across a cohort. **Guided Prompts** — 7 pre-built prompts to help you get started: * Refine patient stratification * Visualize whole slide images * Assess biomarker potential * Export quantitative data * Explore Pathology Explorer capabilities * Visualize tiles with predictions * Get cohort-level statistics Improvements: * Feature descriptions now link to detailed documentation. * Better timeout handling for large slide operations. * 12 MCP tools total (up from 9) #### 2025-12-30 **PDF report export** Reports can now be downloaded as PDF files for easy sharing and archiving. **Bulk vs. spatial transcriptomics concordance** New visualization compares bulk RNA-seq with spatial transcriptomics expression for the same genes. **Gene expression stratification in survival analysis** Kaplan-Meier plots now support patient stratification by bulk gene expression levels (high vs. low expressors). **Resolved visualization timeouts** Gantt visualizations no longer time out on large datasets. #### 2025-12-17 **MOSAIC V5 single-cell data** Added V5 single-cell RNA-seq for colorectal cancer with improved UMAP dimensionality reduction computed at the indication level. **Enhanced visualization clarity** Sankey and Gantt plots now include legends with consistent color coding. Sankey plots auto-prune less relevant pathways. Single-cell plots have improved colorbar layouts. **Improved spatial and single-cell plot accuracy** Updated plot descriptions improve how K Pro interprets analysis requests. **Fixed Sankey plot grouping** Corrected patient grouping errors in treatment pathway visualizations. **Fixed single-cell annotations** Resolved cell type mapping errors. #### 2025-12-05 **Treatment pathway visualization** New Sankey diagrams show patient flows across lines of therapy with transitions and clinical responses. **Corrected volcano plot captions** Captions now correctly display gene/protein counts instead of sample counts. #### 2025-12-01 **Pathology MCP ExplorerInitial release.** * 9 MCP tools for histomics analysis * Support for 6 cell types across TCGA cohorts * Survival analysis with Kaplan-Meier curves * Slide thumbnail visualization * Histomics data download via presigned URLs ### September 2025 #### 2025-09-16 **Cell expression percentages in spatial analysis** Spatial transcriptomics now shows the percentage of cells expressing a target per patient, broken down by cell type. **Improved platform capability guidance** K Pro now more accurately describes its own capabilities when asked. #### 2025-09-02 **In-platform feedback button** Added "Contact us" button for direct feedback and support requests. **Resolved multi-omics display errors** Fixed visualization rendering issues. ### August 2025 #### 2025-08-07 **MOSAIC datasets in data explorer** MOSAIC datasets now appear on the Explore Data page for easier discovery. **Clearer product identification** UI now clearly indicates which product you are using. **Improved multi-omics grouping** Fixed cohort comparison reliability. ### July 2025 #### 2025-07-30 **Explore Data tab** New interface for browsing available datasets (TCGA, MOSAIC Window) before starting analyses. **Advanced filtering and grouping** Filter on any indication-specific clinical variable. Group patients by histomics variables such as cell density from H\&E images. **MOSAIC V5.3 release** Added clinical variables and manual annotations for the full MOSAIC cohort. #### 2025-07-12 **Sample-level multi-omics analysis** Plots can now be generated at the sample level to reveal intra-patient heterogeneity. #### 2025-07-03 **Gene signature support across all plot types** All visualizations now support custom multi-gene signatures. Define pathway-level scores (e.g., `[PTGS2, PTGES, PTGES2, PTGES3]`) and analyze combined expression. ### June 2025 #### 2025-06-20 **Single-cell co-expression analysis** New dotplot visualizes gene pair co-expression at single-cell resolution with Jaccard index scoring. **Chat session deletion** Delete individual sessions for better organization. **Gene signature support (initial release)** Bulk violin, stratified violin, and UMAP plots now support gene signatures. **Suggested prompts** Landing page now shows example queries for faster onboarding. **Fixed bulk RNA-seq calculations** Corrected log2(TPM+1) transformation errors. **Improved population summaries** Resolved demographic table display issues. Questions or feedback? Contact --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.owkin.com/terms-and-support/support/changelog.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.