One regulatory story: how TP53 loss can leave CDKN1A (p21) unactivated, and five data modalities that triangulate it across scales.
Real genomics rarely uses one visualization at a time. Investigating a candidate oncogenic event pulls in population studies, tumor cohort mutation maps, epigenomic signals, pathway topologies, and interaction databases. No single plot answers the question; the triangulation does.
This page takes TP53's loss of canonical activation of CDKN1A (p21) as a running example, viewed five ways. The datasets are synthetic so the same story is visible in every panel; the biological relationships are simplified from the DNA damage response.
Population genetics starts the zoom: before asking about any one tumor, does inherited variation near our gene of interest associate with disease risk? A genome-wide association study (GWAS) tests millions of variants in parallel. The Manhattan plot lays each variant out by its genomic position (x) and statistical evidence (\(-\log_{10}(p)\), y). Loci that rise far above the noise floor are candidate risk regions, not causal genes by themselves.
Figure 1. Genome-wide scan across 22 autosomes. Drag the threshold. The two named peaks sit near TP53 (chr17) and CDKN1A (chr6). Hover any SNP for its position and p-value. Synthetic data.
Two synthetic loci rise above the significance line: chromosome 17 near TP53, chromosome 6 near CDKN1A. GWAS hits are statistical, not mechanistic. They say where to look, not what is broken.
Each tumor carries its own cocktail of somatic alterations: point mutations, truncations, copy-number gains and losses. The oncoprint compresses a cohort into a single grid: columns are patients, rows are genes, colored marks are alteration events. Sort columns by TP53 status to expose co-occurrence patterns.
Figure 2. 80 synthetic patients × 9 genes. Sorted rows show gene alteration frequency on the right. Hover a cell for patient + gene + alteration type. Sorting on TP53 exposes co-occurrence: MDM2 amplifications cluster in the TP53-wildtype stripe. Amplified MDM2 is another way to suppress p53 function.
The oncoprint shows frequency: TP53 alterations are common and tend to be mutually exclusive with MDM2 amplification, a convergent-inactivation signature. CDKN1A itself is rarely altered here, which is part of the point. A mutation grid cannot tell us whether downstream p21 activation actually happens in those patients.
Mutation is not the only way to lower a gene's output. Even in tumors with intact TP53, methylation near the CDKN1A promoter can be associated with reduced transcriptional response. The β value at each CpG site reports methylation level on a 0-1 scale; zoom into the promoter and compare groups of patients.
Figure 3. Top: CpG-level β values across 1.5kb of the CDKN1A promoter. Dots are individual CpGs (hover for position and β). The TSS sits at 0. Bottom: per-sample heatmap — rows are patients, columns are CpGs — revealing that not every sample in the "methylated" group silences the entire promoter.
Methylation can reveal an alternative route to the same readout: a patient with a methylated, low-response CDKN1A promoter may look mutation-normal in the oncoprint but still fail to produce a p21 arrest signal.
The first three views are quantitative; the fourth is structural. The pathway diagram encodes activation and inhibition relationships from decades of biochemistry. Trigger DNA damage to watch the signal propagate and see which branches are lost when TP53 cannot activate its targets.
Figure 4. The canonical TP53 response. Activating edges are arrows; inhibitory edges end in bars. The mutant toggle disables TP53's transcriptional activity in this simplified model. R175H can also have gain-of-function effects that are outside this diagram. Click any node for a one-line description.
If TP53 cannot bind and activate its target promoters, DNA damage fails to produce the p21 cell-cycle arrest signal. The oncoprint shows the lesion is frequent; the pathway shows why it matters.
The curated pathway is a clean story. The protein–protein interaction (PPI) neighborhood is messier: TP53 has dozens of direct physical partners, and its first-degree network reveals how much collateral wiring depends on it.
Figure 5. Drag any protein. Node size scales with degree. Edge thickness scales with synthetic evidence weight. TP53 and CDKN1A are highlighted as the anchors. "Axis path" follows the p53-p21-CDK-RB route from transcriptional response toward cell-cycle control; this suggests places to inspect, not proof of regulation.
No single view settles how TP53 loss drives disease. The Manhattan plot names regions without mechanism. The oncoprint counts events without measuring output. Methylation catches regulatory state invisible to mutation calling. The pathway shows a functional consequence. The PPI shows collateral wiring. Convergent evidence across all five is what makes a claim defensible.