# Multiple Sequence Alignment **Domain:** Bioinformatics / Evolution / Protein Structure **Doc Type:** Canonical Method Node **Maturity:** Foundational ## Definition A **multiple sequence alignment** arranges three or more DNA, RNA or protein sequences to propose homologous positions inherited through common ancestry. ## Evolutionary Signal Alignments make conservation and covariation visible. Positions that remain stable may be functionally constrained; correlated changes can indicate structural or functional interaction. The signal depends on phylogenetic diversity, alignment quality and correct homology assumptions. [[articles/The Evolutionary Roots of Silicon Valley|The Evolutionary Roots of Silicon Valley]] emphasizes that protein-structure prediction can exploit this inherited record. The model is not learning structure from sequence in a historical vacuum; evolution performed a vast distributed experiment whose surviving variants carry information. ## Continuity Context An alignment is a map of relations among descendants, not a restoration of an ancestor. It supports reconstruction and provenance while leaving identity separate. ## Sources / Provenance - Durbin et al., _Biological Sequence Analysis_, 1998. - AlphaFold research overview: https://www.ebi.ac.uk/training/online/courses/alphafold/ ## See Also [[wiki/Comparative Genomics|Comparative Genomics]], [[wiki/Molecular Evolution|Molecular Evolution]], [[wiki/Protein Structure Prediction|Protein Structure Prediction]], [[wiki/Proof of Descent|Proof of Descent]]