1. Getting Started

(1) Website Introduction

EVOKE (Evolutionary Vector Of Key Events) is a comprehensive bioinformatics analysis platform designed to explore the key molecular mechanisms and dynamic evolutionary features in the inflammation-to-cancer transformation process. It aims to systematically characterize molecular events at single-cell resolution during the progression of various cancers.

The EVOKE platform integrates 7,147,991 single-cell data points from 12 cancer types in humans and mice, covering 8 evolutionary vectors (spanning healthy tissue, inflammation, primary tumor, and metastasis stages). Through multidimensional data mining and visualization functions, EVOKE is dedicated to revealing:

  1. Evolutionary trajectories between inflammation and cancer
  2. Dynamic changes in cell composition and functional states across stages
  3. Dynamic regulatory mechanisms of key genes, signaling pathways, and molecular networks in inflammation-to-cancer transformation
  4. Cellular heterogeneity in the tumor microenvironment and its impact on disease progression
  5. Associations between gene expression patterns and cellular trajectories across stages
  6. Biological foundations for potential therapeutic targets and personalized treatment strategies

EVOKE not only provides standardized bioinformatics analysis workflows but also features an interactive visualization interface, enabling researchers to more conveniently gain in-depth insights into the molecular landscape and cellular ecosystem of the inflammation-to-cancer transformation process.

Figure 1: EVOKE Platform Overview

(2) Website Structure and Components

EVOKE is designed with the following six core modules:

Home: The platform's entry point, offering an overview of the project background, data summary, key features, and latest updates to help users quickly grasp EVOKE's core offerings.
Browse: Enables data exploration by cancer type or evolutionary vector, allowing users to gain a comprehensive understanding of dynamic changes in cell composition, trajectory evolution, and gene expression across different pathological stages.
Search: Supports four types of search combinations—gene, cancer type, evolutionary vector, and cell type—enabling users to swiftly locate target genes, cell types, samples, or analysis results to meet precise query needs.
Download: Centralizes access to key data files, including cancer progression-specific marker genes, cross-stage consistently altered genes, functional pathways, and trajectory changes. It supports multi-condition filtering and batch downloads, facilitating local analysis by users.
Tools: Integrates lightweight online visualization tools that allow users to explore microenvironment changes across cancers, stages, patients, and genes—including expression patterns and pathway enrichment variations. This module supports preliminary exploration and hypothesis validation.
Help: Provides detailed user guides, module functionality explanations, and FAQs, along with open channels for technical support and feedback, ensuring users can seamlessly acquire and analyze data.

2. Functionality Guide

(1) HOME

The EVOKE data platform comprehensively integrates 7,147,991 single-cell data points, covering 12 cancer types from humans and mice across 8 evolutionary vectors (spanning healthy tissue, inflammation, primary tumor, and metastasis stages).

The homepage offers a convenient quick-search feature, enabling users to efficiently query and explore target data.

Figure 2: EVOKE Homepage Overview
Figure 3: Quick Search Feature

Additionally, the homepage features an interactive word cloud highlighting key information across the four stages of inflammation-to-cancer transformation. Clicking on a keyword in the word cloud directs users to the corresponding search page, which displays the relevant search results.

Figure 4: Interactive Word Cloud

(2) BROWSE

The Browse page is a key module in the EVOKE platform, designed for systematic data exploration by Cancer Type or Progression Type. Users can quickly select specific cancers or disease progression stages via intuitive dropdown menus to delve into associated cellular ecosystems, trajectory patterns, and key gene regulatory information.

The Browse page adopts a dual-column layout:

  • Left Column (Menu Bar): Categorized by human and mouse data, each species includes two units (Cancer Type and Progression Type), with each unit supporting a three-level dropdown menu for selection.
  • Right Column (Display Area): A scrollable page presenting analysis charts and explanatory information for the selected content, organized into multiple functional sections with anchor links for easy navigation.
Figure 5: BROWSE Module Interface

When users first enter the main page, they get an overall view of the data used in the EVOKE platform and the results from our analysis.

On the left, a scrolling bar chart shows the main cell types at four different stages: healthy, inflammation, primary tumor, and metastasis. The animation helps users clearly see how the tumor microenvironment changes as cancer progresses.

In the center, a word cloud highlights the most frequently mentioned terms from our analysis, giving users a quick sense of the key information.

Figure 6: Data Overview Visualization

At the bottom, another bar chart focuses on the top four dominant cell types and shows how their proportions change across the four stages. This helps explain why it's important to study how the microenvironment evolves during cancer development.

Figure 7: Cell Proportion Changes

① Browse by Cancer Type

Selecting Cancer Type in the left menu reveals a list of specific cancers, such as Breast Invasive Carcinoma, Clear Cell Renal Cell Carcinoma, and Colorectal Adenocarcinoma. Each cancer type is further subdivided into disease progression stages: Inflammation, Tumorigenesis, Malignant Transformation, and Metastasis. Clicking a cancer name directs users to its dedicated browse page, displaying data and charts for all stages.

The page is organized into three main sections:

Section Subsection Description
Cell Section Tumor Microenvironment Visualizes the composition and changes in the tumor microenvironment across stages.
Cell Proportion Displays the proportion of cell types and their changes across progression stages.
Impact of Progression Analyzes the influence of disease progression on cellular ecosystems.
Trajectory Section Trajectory of Progression Shows developmental trajectories of cell states across stages.
Trajectory Information Details the foundational data and methods used for trajectory construction.
T-Impact of Progression Analyzes the impact of different stages on trajectory direction and speed.
Trajectory Correlation Compares trajectory similarities across cancer types or stages.
Gene Section Key Genes of Progression Identifies genes critical to the evolutionary process.
Gene Network Analyzes regulatory networks and modules among genes.
Prognostic Impact Evaluates the association of key genes with patient prognosis.
Function Regulation Displays functional enrichment and pathway regulation patterns.

Clicking any submodule in the left menu automatically scrolls the right display to the corresponding charts and explanations, enabling users to focus on their areas of interest.

Figure 8: Cancer Type Browse Example

② Browse by Progression Type

Selecting Progression Type in the left menu allows users to view data across multiple cancer types by stage, with two perspectives:

  • Multi-Stage: Covers four main evolutionary stages—Inflammation (Health → Inflamed), Tumorigenesis (Health → Primary), Malignant Transformation (Inflamed → Primary), and Cancer Metastasis (Primary → Metastasis). Selecting a stage (e.g., Tumorigenesis) leads to a unified page integrating cell composition, trajectory changes, and gene characteristics across all cancers for that stage. The page structure mirrors the Cancer Type layout (Cell, Trajectory, Gene sections) but aggregates data from multiple cancers.
  • Single-Stage: Offers a sample-based perspective (Health, Inflamed, Primary, Metastasis), ideal for exploring the overall landscape and characteristics of a specific tissue state from a clinical or histological viewpoint.
Figure 9: Progression Type Browse Example

(2) SEARCH

The Search page is a core module enabling multidimensional queries to quickly locate data related to specific genes, cell types, cancer types, or disease stages. It is divided into two sections: Quick Search and Advanced Search, catering to exploratory and precise queries, respectively.

Quick Search
A straightforward single-line input box for keyword-based fuzzy searches. Users can enter terms like gene names, cancer type keywords, or cell type names. Below the input box, Example Searches guide users on typical inputs and formats. Upon clicking search, the system directs to a results page displaying all relevant charts and data.
Advanced Search
Supports precise and combinatorial queries with four conditions: Gene Symbol, Cell Type, Cancer Type, and Progression Stage. Users can enter any gene name in the Gene Symbol field, with the system offering fuzzy matching and auto-completion for common or high-interest genes. Cell Type, Cancer Type, and Progression Stage are predefined categories aligned with database annotations, enabling standardized searches. For example, users can select a specific cancer (e.g., Lung Adenocarcinoma), stage (e.g., Tumorigenesis), or cell subtype (e.g., CD8+ T cell) to narrow their search. Conditions can be used individually or combined to support queries like multi-cancer analysis of a single gene across stages or key regulatory genes in a specific cancer stage and cell type. Clicking the search button generates a results page structured similarly to the Browse page, providing a comprehensive view of the target data across disease contexts.
Figure 10: SEARCH Module Interface

(3) TOOLS

① Progression Cancers
This tool identifies oncogenic hotspots for target genes in pan-cancer single-cell data. By integrating single-cell resolution expression data, it pinpoints genes acting as Cross-stage consistently altered genes in specific cancer types and reveals conserved mechanisms in molecular events (e.g., gene regulation) and disease contexts. The visualized outputs help users intuitively understand a gene's role in cancer progression.

Users first select a marker type (ProgressionType for progression-specific markers or Cellmarker for classic cancer markers), then choose a suggested gene (e.g., A2M, AIF1), cell type (e.g., Cancer Cells), and sample type. The tool displays the marker's expression patterns across cancer types (e.g., BRCA, CCRCC) and progression stages (Health, Inflammation, Primary Tumor, Metastasis).
Figure 11: Progression Cancers Tool
② Progression Pathway Analysis
This tool analyzes user-input gene sets to uncover their associations with cancer progression hallmarks, such as invasion, metastasis, and microenvironment remodeling. Leveraging functional enrichment data, it identifies synergistic pathway hubs (e.g., epithelial-mesenchymal transition [EMT], immune evasion, metabolic reprogramming) to reveal the biological "scripts" driving malignant tumor development.

Users input a gene set, and the tool maps these genes to progression-related pathways using databases like KEGG or GO, aiding analysis of their functional relationships.
Figure 12: Pathway Analysis Tool
③ Progression Annotation
This tool dissects tumor microenvironment (TME) changes across cancer contexts, identifying key cell types driving tumor progression. It visualizes trends in cell population proportions across stages (e.g., Health, Inflammation, Metastasis) and dives into patient-specific microenvironment states, highlighting potential therapeutic targets (e.g., immune evasion-related cells).

Users select patients from different cancers, and the tool returns dynamic changes in their TME during cancer progression.
Figure 13: Progression Annotation Tool
④ Progression Gene
This tool identifies key genes driving tumor progression and their roles in the TME. Users select a Tumor Type, then a key gene, and the tool returns its contributions to microenvironment remodeling or tumor progression.
Figure 14: Progression Gene Tool
⑤ Progression Function
This tool explores the roles of functional programs (e.g., GO, KEGG, Hallmark gene sets) in tumor progression. Users input a biological process (e.g., EMT, angiogenesis), and the tool analyzes its activation patterns across specific cancer types and stages using functional enrichment data.
Figure 15: Progression Function Tool

(4) DOWNLOAD

The platform provides key file downloads, divided into two sections:

Human Data includes four categories:
• Dynamic_Cell_Marker_Repository: Genes specifically expressed at different cancer progression stages.
• Cross-stage consistently altered genes: Consistently altered genes across progression stages.
• Cellular_Trajectory_Map: Summarized data on cell differentiation dynamics during cancer progression.
• Tumor_progression_function: Progression-related functions identified during cancer progression.
Mouse Data includes three categories:
• Dynamic_Cell_Marker_Repository: Genes specifically expressed at different cancer progression stages.
• Cross-stage consistently altered genes: Consistently altered genes across progression stages.
• Cellular_Trajectory_Map: Summarized data on cell differentiation dynamics during cancer progression.

Clicking a category directs users to a detailed download page, where they can select individual cancer or pan-cancer data tables for download.

Figure 16: DOWNLOAD Module Interface

3. Chart Interpretation

Tumor microenvironment

The Tumor Microenvironment chart displays the activity profiles of various cell types within the selected cancer. Hovering over a data point in the chart reveals detailed information specific to that point, providing insights into the cellular dynamics and interactions within the tumor microenvironment.

Figure 17: Tumor Microenvironment Visualization

Cell proportion

The Cell Proportion chart illustrates the dynamic changes in the infiltration of various cell types across different sample stages of the target cancer. Additionally, we performed further clustering of these cell types, as shown in the rightmost panel labeled Progression Cluster. This allows users to assess whether certain cell types exhibit similar infiltration patterns during cancer progression.

Figure 18: Cell Proportion Visualization

Users may also choose to display individual cell types within the target cancer, allowing for a focused examination of their infiltration dynamics across various stages of progression.

Figure 19: Individual Cell Type Progression

Impact of progression

Shifting Tissue Preference in Tumor Progression

This section provides a detailed illustration of how various cell types exhibit shifting preferences across different stages of disease progression in the selected cancer. It helps users gain deeper insights into how the cellular ecosystem drives cancer evolution and uncovers potential underlying biological mechanisms.

Figure 20: Shifting Tissue Preferences

Progression-Associated Disturbance for Each Cell Type

This section illustrates the dynamic perturbations of individual cell types throughout cancer progression, analyzed at the level of each patient. By hovering over different data points, users can view detailed information about a specific cell type from a specific patient. As shown in the example below for CRC (colorectal cancer), compared to the Tumorigenesis stage, patients are more broadly distributed during the Inflammation and Cancer Metastasis stages. This indicates greater heterogeneity and more pronounced perturbations in these stages.

Figure 21: Cell Type Disturbance Analysis

Trajectory of progression

This section presents all progression trajectories identified in the selected cancer type. The table displays the starting and ending states of each trajectory, along with their associated driver genes and the top five most significantly expressed key genes during the transition.

By clicking on any Trajectory ID in the table, users can view more detailed information below, including the dynamic changes in cell states along the selected trajectory and the specific patients associated with it.

Figure 22: Trajectory Table Overview
Figure 23: Trajectory Detail View

When hovering over specific positions in the trajectory plot, the interactive chart will automatically display detailed information about the indicated cell, including its cell type, pseudotime point, and corresponding sample stage.

As shown in the figure below, for the CRC_TR001 trajectory, the progression starts from ENS glia cells and ends with macrophages.

Figure 24: Interactive Trajectory Visualization

In addition, we provide an interactive plot that dynamically illustrates the proportional changes of different sample stages along the pseudotime trajectory. This allows users to intuitively observe how cell states shift throughout the course of cancer progression.

Figure 25: Pseudotime Progression Analysis

Trajectory Correlation

In this section, we visualize the similarity relationships among progression trajectories within the selected cancer type. By clicking on any pathway in the network graph, the panel on the right will display other trajectories related to the selected one. This helps users better understand the dynamic changes of trajectories across different sample stages and progression phases.

Figure 26: Trajectory Correlation Network

Marker genes in progression And Key genes of progression

The "Marker Genes in Progression" table displays cell type-specific marker genes expressed in different sample types within the selected cancer. In contrast, the "Key Genes of Progression" table highlights the driver genes associated with various progression stages during cancer development.

When clicking on a marker gene in the table, the panel below will automatically display its interaction network as well as its dynamic expression pattern along the trajectory.

As shown in the figure below, the plot illustrates how the expression levels of the driver genes HLA-DPA1, HLA-DQB1, and HLA-DRA change throughout cancer progression.

Figure 27: Marker Genes Table
Figure 28: Gene Interaction Network
Figure 29: Gene Expression Dynamics

Prognostic Impact

This section presents key genes associated with prognosis in the selected cancer type. Interactive visualizations, including forest plots and survival curves, are provided to illustrate the prognostic relevance of these genes.

When a gene is selected from the table, the corresponding dynamic plots on the right and below will update automatically to reflect its prognostic impact.

Figure 30: Prognostic Gene Analysis
Figure 31: Survival Curve Visualization

Functional Regulation

This section visualizes the dominant biological functions across different stages of cancer progression, helping users intuitively understand the dynamic shifts in functional activity that drive tumor evolution.

Figure 32: Functional Activity Analysis
Figure 33: Pathway Regulation Patterns

4. Contact Us

EVOKE Development Team

Yunpeng Zhang: zhangyp@hrbmu.edu.cn

Xia Li: lixia@hrbmu.edu.cn

Jing Bai: baijing@hrbmu.edu.cn

College of Bioinformatics Science and Technology

Harbin Medical University 194 Xuefu Road, Harbin 150081, China

Phone: 86-451-86615922

Fax: 86-451-86615922

Email: zhangyp@hrbmu.edu.cn