comparative genomics

Imagine two ancient manuscripts copied by hand from the same lost original over many centuries. Where every surviving copy keeps an identical phrase, that phrase was probably important — change it and the meaning broke, so scribes who altered it left no readable descendants. Where copies differ freely, that text mattered less. Comparative genomics reads genomes the same way: by lining up the DNA of different species, what stayed the same and what drifted apart tells you about function and history.

The core method is to align genomes from related species and look at the pattern of conservation. A stretch of DNA that has barely changed across hundreds of millions of years of separate evolution is almost certainly doing something essential, because mutations there were weeded out by selection — these are conserved elements, and conservation is one of the strongest signals of function we have, often flagging regulatory regions that no gene-prediction program would catch. Comparison also sorts related genes into orthologs (the 'same' gene in two species, descended from a common ancestral gene) and paralogs (copies that arose by duplication within a lineage), and reveals synteny, blocks where gene order is preserved across species, betraying their shared ancestry.

This is how we read function and deep history straight out of sequence. Comparing the human and mouse genomes, or human and chimpanzee, pinpoints the few percent of the genome under tight constraint and the rare regions that changed specifically in our lineage. The honest limits are real, though: high conservation strongly suggests function but does not prove it, low conservation does not prove uselessness (some functional elements evolve fast), and orthology calls can be wrong when genes are gained, lost, or duplicated, so a careless ortholog assignment can carry a false functional inference along with it.