An editorial synthesis that gathers a gene, its target, and their evidence into one readable page.
A real-world read on a gene, such as a GeneCards entry, the first slide of a collaborator's deck, or the figure that opens a review, gathers prior charts into one page. The Gene Portrait is that view: six mini-panels inheriting vocabulary from the standalone techniques, arranged so the whole story reads in under a minute. It is an editorial synthesis, not a clinical decision-support tool.
The portrait below is built around the TP53 to CDKN1A axis. TP53 encodes p53, a stress-responsive transcription factor. CDKN1A encodes p21, a cyclin-dependent kinase inhibitor whose expression is tightly controlled by p53 and helps mediate G1 arrest. The same synthetic patients drive every panel, sliced differently. When the panels move together, the story converges.
Inherited from the lollipop (09). Mutant cohort shows the three canonical hotspots at residues 175, 248, and 273, all in the DNA-binding domain.
Inherited from the MA, volcano and waterfall plots. Distribution of CDKN1A expression across patients; dashed line marks the cohort median.
Inherited from the five views epigenome panel. Per-CpG mean β value across the CDKN1A promoter. Higher β means more methylated at that CpG, not automatically silent.
Inherited from the pathway panel of the five-views article. A broken edge is dashed red. Downstream outcomes show which branches reach completion.
Inherited from the survival curves (07). Curves show the two cohort arms; the log-rank p-value appears beside the legend.
Inherited from the forest plot (14). Each row is a published study; the summary diamond pools effect sizes with inverse-variance weighting.
Figure 1. Hover any panel for detailed values. The same two cohort buttons drive every panel simultaneously. The labelled circled numbers trace back to the earlier article that introduced each visual primitive.
The six panels share a cohort axis. Each is an independent test of the same hypothesis; the portrait shows whether evidence converges.
For the mutant cohort: the variant map lights up at hotspot positions, CDKN1A expression shifts down, promoter methylation is higher in this synthetic cohort, the p53-to-target edges weaken, survival separates downward, and the pooled meta-analytic effect is consistent. Each panel alone could be noise. Together they make a coherent case, convergent but not conclusive, that the p53 response is impaired.
A portrait is specific to a claim. This one is built for "TP53 alteration weakens p53-responsive CDKN1A induction and is associated with worse outcomes." A different question, such as "MDM2 amplification as an alternative route," needs its own portrait with a different cohort axis.
| Dimension | Metric | Wildtype | Mutant | Δ |
|---|---|---|---|---|
| Variants | Hotspot mutation rate | 0% | 82% | +82 pp |
| Expression | CDKN1A median log₂FC | +1.8 | −2.1 | −3.9 |
| Methylation | Promoter mean β | 0.14 | 0.36 | +0.22 |
| Signaling | TP53 to CDKN1A activation | active | weak | context |
| Survival | Median OS (months) | 42 | 21 | −21 |
| Meta-analysis | Pooled log₂ HR (95% CI) | — | 0.94 (0.71–1.18) | — |
Figure 2. The same evidence as the portrait above, in numeric form. A condensed table like this is what tends to get cited and re-circulated alongside the figure. Synthetic data — not clinical guidance.
The full visual vocabulary covered: heatmaps and dendrograms for raw matrices; MA and volcano plots for statistical filtering; UMAP for high-dimensional landscapes; genome browsers and circos plots for spatial and long-range context; survival curves, dot plots, and lollipops for clinical and protein-level reads; a triangulation article walking one story through five modalities; ChIP-seq for physical binding; multi-omics heatmaps and waterfalls for layer integration; forest plots for meta-analytic rigor; and this portrait for single-page synthesis.
cBioPortal, GeneCards, TCGA's PanCancer Atlas viewers, in-house lab portals — all are elaborations on the same primitives.