Cortex Viz

From analysis to insight, in one conversation

Cortex builds interactive, presentation-ready reports from your analysis results. Ask for what you need, refine until it's right, share with a link.

Breast Cancer Ancestry Analysis: Report

Report Sections

Summary
Study Design
Cohort Analysis
Differential Expression
Pathways
Dose-Response
Immune
TF Activity
Networks
Genomics
Survival
Synthesis
Limitations

Generated by Inflexa · Feb 21, 2026

Differential Expression

Ancestry-stratified gene expression

BRCA1TP53
Ask Cortex to refine this report...Send
From Chat to Report

Your analysis is done. Now present it.

After running differential expression, pathway enrichment, immune profiling, and survival analysis, you're sitting on results from multiple runs. One message turns them into a structured report.

Cortex knows what you ran. It pulls the figures, tables, and statistics from your analysis history and assembles them into presentation-ready sections.

One message. Every figure traced to its source run.

Cortex · Breast Cancer Ancestry Analysis5 runs completed
Analysis session · TCGA-BRCA + RA-QA · 1,084 samples

Completed Runs

Differential ExpressionPathway EnrichmentImmune DeconvolutionTF ActivitySurvival Analysis
Build a report from these results. Focus on how genetic ancestry shapes breast cancer biology: the DE patterns, immune differences, and survival outcomes across EA, AA, and Arab cohorts.
CORTEX

I'll build a comprehensive report from your 5 analysis runs. Here's the structure:

Report Outline

1. Summary2. Study Design3. Cohort Analysis4. Differential Expression5. Pathways6. Dose-Response7. Immune Profiling8. TF Activity9. Networks10. Genomics11. Survival12. Synthesis13. Limitations

Generating now: pulling figures from your DE, pathway, immune, TF, and survival runs...

Done
R

Breast Cancer Ancestry Analysis: Report

13 sections · 23 figures · 8 tables · Click to open

Summary

The whole story on one screen

Every report opens with a structured summary: cohort size, total DEGs, key pathway hits, and immune signatures, all quantified and sourced to the underlying runs.

Top findings are ranked by effect size and statistical significance, not just listed. Reviewers see what matters first.

A 30-second read that replaces a 30-minute deep dive.

Report Summarycomplete

Study Overview

1,084
samples
4,217
DEGs (FDR < 0.05)
3
ancestry cohorts
47
enriched pathways
12
immune cell types
p=0.003
survival (AA vs EA)

Key Findings

PI3K/Akt/mTOR pathway dominates African American tumors

SHAP importance 0.510 · NES = 2.34, FDR = 1.2e-06

Th2 immune infiltration elevated 2.3-fold in AA vs EA

Immune deconvolution · p = 4.1e-04

NT5E and estrogen response drive European ancestry classification

SHAP importance 0.234, 0.134 · Ancestry-stratified ML model

Significant survival divergence between ancestry groups

Log-rank p = 0.003 · AA 5-year OS: 72% vs EA: 84%

Sourced from 5 analysis runs · TCGA-BRCA + RA-QA
Differential Expression

Every gene, in context

Interactive volcano plots render directly in the report. Hover any point to see the gene name, fold change, and adjusted p-value. Click to jump to the gene's full profile.

Cortex stratifies the analysis by ancestry group, revealing gene expression patterns that are invisible in a pooled analysis. BRCA1 shows 3.2-fold upregulation in AA samples: a signal lost when cohorts are combined.

4,217 DEGs across 3 ancestry cohorts: each plotted, searchable, and linked to downstream analyses.

Volcano Plot · AA vs EAinteractive
AA vs EAArab vs EAAA vs Arab|log2FC| > 1 · FDR < 0.05
-log10(p-adj)log2 Fold Change036912151821-3-2-10123BRCA1EGFRMYCKRASTP53ESR1NT5EFOXA1
Upregulated (1,892) Downregulated (2,325) NS
Hover to inspect · Click to explore
Survival Analysis

Outcomes that tell the full story

Kaplan-Meier curves with confidence intervals, risk tables, and log-rank statistics, all generated from your survival run and rendered as interactive charts.

African American patients show significantly lower 5-year overall survival (72%) compared to European American (84%). The divergence becomes apparent after year 2, consistent with known ancestry-linked disparities in treatment access and tumor biology.

Clinical context that no spreadsheet can provide: survival curves stratified by your variables of interest.

Kaplan-Meier · Overall Survival by Ancestryinteractive
Log-rank p = 0.003n = 1,084Median follow-up: 4.2 yr
Survival ProbabilityYears0.000.200.400.600.801.00012345EA (n=734)Arab (n=24)AA (n=326)
EA: 84%Arab: 78%AA: 72%

Number at Risk

Group0 yr1 yr2 yr3 yr4 yr5 yr
EA734698671648629617
Arab242321201919
AA326302278258242235
Iterate

Refine it like you'd brief a colleague

Reports aren't one-shot. Rearrange sections, add context, adjust visuals, all through natural language. Every edit preserves the underlying data and statistical rigor.

Tell Cortex what to change. It understands your analysis, so "add the immune comparison to the summary" doesn't require explaining what that means.

No re-running analyses. No exporting and re-importing. Just say what you want.

Cortex · Report IterationContent Edit
Move the immune infiltration comparison into the summary: that's the headline finding. And drop the dose-response section, we didn't run that analysis.
CORTEX
Done. Two changes applied:
MOVEDTh2 infiltration finding (2.3-fold, p=4.1e-04) added as key finding #2 in Summary
REMOVEDDose-Response section removed. Report now has 12 sections.

Before

1. Summary

2. Study Design

3. Cohort

4. DE

5. Pathways

6. Dose-Response

7. Immune

...

After

1. Summary ★

2. Study Design

3. Cohort

4. DE

5. Pathways

6. Immune

...

Cortex · Report IterationVisual Edit
In the volcano plot, only label genes with |log2FC| > 2. It's too cluttered right now. Also use a more distinct color for the Arab cohort: the amber is hard to see on the light background.
CORTEX
Updated. Here's what changed:
CHARTVolcano labels now filtered to |log2FC| > 2: 23 genes labeled (was 156)
COLORArab cohort color changed from amber to violet across all charts

Before

ESR1CDH1GATA3FOXA1EGFRBRCA1PIK3CAMYCTP53

After

EGFRBRCA1ESR1
Synthesis

Conclusions backed by every upstream analysis

The Synthesis section doesn't just summarize: it integrates. Each conclusion is linked to the specific analyses that support it, with confidence scored by evidence convergence.

When three independent analyses point to the same biological story, Cortex flags it as high-confidence. When evidence is mixed, it says so.

Not a summary of summaries. A synthesis that traces every claim to its evidence chain.

Synthesis · Integrated Conclusions4 conclusions
HIGH CONFIDENCE4 converging analyses

PI3K/Akt/mTOR signaling is the dominant ancestry-linked pathway in AA breast cancer

Supported by differential expression (NES=2.34), SHAP feature importance (0.510), pathway enrichment (FDR=1.2e-06), and TF activity analysis showing upstream AKT1 activation. This represents a potential therapeutic target for ancestry-stratified treatment.

DE: NES=2.34SHAP: 0.510Pathway: FDR=1.2e-06TF: AKT1 activated
HIGH CONFIDENCE3 converging analyses

Immune microenvironment composition diverges significantly by ancestry

Th2 infiltration is elevated 2.3-fold in AA tumors (p=4.1e-04). This is corroborated by LAG3 expression differences (SHAP=0.290) and survival analysis showing immune-stratified outcomes.

Immune: 2.3-fold Th2SHAP: LAG3=0.290Survival: p=0.003
MODERATE2 converging analyses

Estrogen receptor signaling drives EA-specific tumor biology

ESR1 and FOXA1 are top SHAP features for EA classification. Estrogen response hallmark pathway is significantly enriched in EA tumors. However, this may partly reflect subtype composition differences.

SHAP: ESR1=0.134DE: Estrogen resp.
EXPLORATORY1 analysis, small cohort

Arab cohort shows intermediate molecular profile

RA-QA cohort (n=24) clusters between EA and AA on principal components. Limited sample size precludes definitive conclusions: validation with larger Arab-ancestry datasets recommended.

Cohort: n=24⚠ Low power
Share

One link. Full context.

Share your report with a link: collaborators see the same interactive visualizations, the same evidence chains, the same conclusions. No exports, no PDFs, no lost formatting.

Team members or external reviewers open the link and get the full report. Every chart is interactive. Every finding is traceable. The context travels with the work.

Your collaborator sees exactly what you see, not a flattened copy.

Cortex · Share Report
R

Breast Cancer Ancestry Analysis

12 sections · 23 figures · 8 tables · Updated 2 min ago

Final

Share link

https://public.inflexa.ai/33558495/breast-cancer-ancestry

Access

🌐

Anyone with the link

No sign-in required

Can view ▾

Shared with

RT

Radu Tanasa

radu@panomics.bio

Owner
SC

Sarah Chen

s.chen@research.edu

Can view ▾

12 sections. One report. Built from your runs.

Every section generated from analysis results, not templates. Structure adapts to what you ran.

Overview

Summary

Key stats, top findings, study overview

Study Design

Cohorts, comparisons, methods used

Cohort Analysis

Sample demographics, group composition

Analyses

Differential Expression

Volcano plots, DE tables, gene rankings

Pathways

Enrichment scores, pathway networks

Immune Profiling

Cell composition, infiltration estimates

TF Activity

Transcription factor regulation scores

Networks

Pathway interaction graphs

Genomics

Variant analysis, mutation profiles

Conclusions

Survival

KM curves, log-rank tests, risk tables

Synthesis

Integrated conclusions, confidence scores

Limitations

Caveats, power analysis, known gaps

See how your analysis becomes a report

Request a demo with your own data or explore a public analysis.