A genome browser lines up genes, binding sites, conservation and reads on one coordinate axis, so evidence at one stretch of DNA reads together.
Part 1 measured methylation in the promoter of CDKN1A, the series' p21 gene, and Part 4 measures p53 binding to it. Where those sites sit relative to the gene and its neighbors is what a browser shows, as tracks of intervals and values stacked under a ruler of the reference genome.
Positions count bases along a chromosome in two conventions. The browser's location box, VCF and Ensembl count from 1 and include both ends: CDKN1A's first exon is . BED files and UCSC's tables count from 0 and leave the end out: . The start is one lower, the end agrees, and the length is end minus start, bases. Nothing in a bare number says which, and Figure 3 shows what a wrong guess does.
Point at the tracks, or tap them, to read a base in both conventions.
Click, tap or tab to a gene, p53 site, island or repeat for details.
Figure 1. (GRCh38), marked on chromosome 6. Ensembl canonical genes; el-Deiry's two p53 response elements; CpG per 100 bases computed from the sequence, over UCSC's island calls; RepeatMasker repeats; phyloP100way clipped at −4 and 7.5. All real data. Drag sideways to pan; pinch, ctrl+scroll or the buttons zoom; on the focused panel, arrow keys pan and + and − zoom.
Only of the genes in these kb code for protein, CDKN1A and RAB44. Exons are tall where they code and short where they are untranslated; chevrons on the intron line give the strand. PANDAR and DINOL, non-coding RNAs named for their induction by DNA damage, run on the minus strand through CDKN1A's promoter.
Most of the window has about CpGs per 100 bases, far below what its base composition predicts, because methylated cytosines in CpGs tend to mutate to thymine. UCSC calls an island where at least 200 bases are at least half G or C and hold at least 0.6 of the expected CpGs. The main island runs from to around the transcription start, the stretch Part 1 followed. Both p53 response elements sit outside it, and bases upstream.
phyloP, from an alignment of 100 vertebrate genomes, is positive where a base changed less than neutral DNA would have and negative where it changed more. It disagrees with the island: exon 1, inside the island and untranslated, averages , while the coding part of exon 2 averages . The 3′ p53 site averages , nearly all from its second half ( against ; see the "3′ p53 site" view). The most conserved stretch in the window is LAP3P2, a processed pseudogene (a copy of LAP3's mRNA), at over bases. The browser shows that the copy changed little, not why.
Figure 2. CDKN1A's three Ensembl transcripts (release 116) on the genome, and the chosen one spliced into mRNA below at its own scale; bands join each exon to its place in the mRNA. The arrow at each start points the way transcription runs; the marks above are the p53 sites. Click a row or button to choose a transcript.
An exon is not the same as coding sequence. CDKN1A's canonical transcript has three exons, and exon 1 is all 5′ untranslated region: the start codon lies bases into exon 2, and the stop codon ends bases into exon 3, ahead of bases of 3′ UTR.
Of Ensembl's two other transcripts, one adds a non-coding exon and the other starts bases upstream, with the 5′ p53 site inside its first exon: the same 20 bases are promoter in one and exon in the other. "Exon 1" means nothing until the transcript is named, usually the MANE Select one that Ensembl and NCBI agree on.
Signal tracks from sequencing start as reads, 100 to 150 bases each, aligned where they best fit the reference; in repeats, of Figure 1's window, several places fit equally well. Stacked, the reads form a pileup. Many reads disagreeing with the reference the same way suggest a variant; a lone one is more likely a sequencing error. Part 4's ChIP-seq tracks are pileups counted in bins.
Figure 3 is a pileup at TP53 codon 175, the R175H hotspot of Parts 1 and 4. Its simulated reads each have a chance of carrying the mutation and about one base in miscalled. At the mutated base of reads carry T.
Point at or tap a column for its base counts.
Figure 3. TP53 exon 5 around codon 175 (chr17, GRCh38). Bases, codons and phyloP100way are real; the reads are simulated (seeded), and coverage (non-reference share filled) is counted from them. The outlined column is where the chosen reading lands. The transcript view mirrors and complements the sequence so TP53 reads 5′ to 3′.
The same mutation is written two ways: VCF 17 7675088 . C T (1-based position, reference C, new base T) and BED chr17 7675087 7675088. Read the VCF number as a BED start and a program lands one base right; read the BED start as a 1-based position and it lands one base left. Both neighbors are G. A program that checks the reference base stops; one that doesn't finds of reads differing to the right, of to the left, and calls a hotspot mutation absent.
TP53 is on the minus strand, so it runs right to left in the genome view. VCF gives the plus strand, C to T; the clinical name c.524G>A gives the transcript strand, where the base is G. In the transcript view codon 175 reads CGC, arginine, which the mutation makes CAC, histidine. That base scores on phyloP, against for exon 5 as a whole and in the flanking introns.
A browser panel shows one contiguous stretch of one chromosome. IGV and the UCSC browser can split the screen into several regions, and IGV colors a read whose mate aligned to another chromosome, enough for a link or two. To see every link at once, the next part bends the chromosomes into a circle.