[Verse 1] Take a genome, shred it into overlapping little strings, Lay them end to end and let the matching edges sing, That's overlap, layout, consensus — O-L-C by name, Building genomes like a jigsaw missing half its frame. Switch the lens now, chop each read to k-mer bites instead, String the nodes where k-minus-one letters overlap and spread, That's a de Bruijn graph, and if the copies all ran clean, One Eulerian path walks through and shows you what it's seen. But errors twist the picture, bubbles bloom where bases lied, Tips are dead-end alleys — snip them off and let them slide. [Chorus] O-L-C, then de Bruijn, Euler ties the thread, Trim the tips and pop the bubbles, chase the graph ahead, Repeats are the ceiling, read length breaks the wall, Count your k-mers, sketch your hash — that's how we cover all. [Verse 2] Repeats are the palimpsest — a page erased and written twice, same sequence laid on old sequence, no seam to tell the price, Short reads stumble in that overlap, lost inside the maze, SPAdes and MEGAHIT dig short contigs from the haze, Long reads stretch clean across the trap like a bridge above a creek, Hifiasm, Flye, and Verkko chase the telomere's far peak, Haplotypes unbraided now, both parents' strands laid bare, Telomere to telomere — the whole map, end to end, no tear. [Verse 3] Hi-C throws a net of contacts, folds the draft into shape, Optical maps mark the mileposts, scaffolds start to take, Polish every letter twice till the consensus base is sure, Then measure what you've built — is the finished genome pure, N50's the median brick, NG50 corrects for scope, auNG draws the whole curve out — the area under hope, BUSCO counts the genes you'd bet on always keeping near, Merqury grades your QV, QUAST hunts trouble in the gear. [Chorus] O-L-C, then de Bruijn, Euler ties the thread, Trim the tips and pop the bubbles, chase the graph ahead, Repeats are the ceiling, read length breaks the wall, Count your k-mers, sketch your hash — that's how we cover all. [Verse 4] Jellyfish and KMC, they're counting every k-mer chord, Plot the spectrum, find the peaks, let GenomeScope afford A guess at genome size and how heterozygous you run — Two peaks mean two haplotypes, one lonely peak means one. Bloom filters hold a rumor, space-cheap, might swear "it's there," Count-min sketch remembers tallies without the full affair, MinHash grabs the smallest hashes, FracMinHash keeps a share, Mash and sourmash cluster genomes — ANI tells who's kin to spare. [Bridge] Mask the repeats first so genes don't hide behind the weeds, AUGUSTUS and Prodigal predict from raw ab initio seeds, BRAKER blends the evidence, RNA proof and homology, Function tags get stitched on last — that's the whole biology. [Chorus] O-L-C, then de Bruijn, Euler ties the thread, Trim the tips and pop the bubbles, chase the graph ahead, Repeats are the ceiling, read length breaks the wall, Count your k-mers, sketch your hash — that's how we cover all.
← Module 6 — Sequence Algorithms I: Alignment (Weeks 10–11) | Module 8 — Variant Discovery & Genotyping (Weeks 14–16) →