E-Value (BLAST)
Definition
The expected number of alignments as good as the query result that would occur by chance in a database of the given size. An E-value of 1e-100 means one such hit is expected in 10^100 random searches. For forensic species identification, E-values below 1e-50 with identity above 98% are typically considered sufficient for a confident species call.
- Meaning
- Expected number of chance alignments as good as the hit
- Interpretation
- Lower value = less likely to be chance
- Forensic confidence threshold
- Below 1e-50 with identity above 98%
- Depends on
- Database size searched
Common questions
Why is a very low E-value not proof of species identity by itself?+
The E-value only measures how unlikely the alignment score is by chance given the database size; it says nothing about database completeness or whether closely related species are even represented. A confident call also requires checking percent identity and query coverage.
How does E-value change if the same search is run against a larger database?+
It increases, because a larger database gives more opportunities for a chance alignment of equal quality. Comparing E-values across searches only makes sense when the database size is the same or explicitly accounted for.
Related terms
- EMPOP SAM2
- The submission quality-assurance module used by the EMPOP database to screen incoming mtDNA sequences for phylogenetic inconsistencies, sequencing artefacts, and nomenclature errors....
- Haplogroup
- A monophyletic branch of the human mitochondrial (or Y-chromosome) phylogenetic tree defined by a set of derived mutations. Haplogroup assignment is the...
- INSDC
- International Nucleotide Sequence Database Collaboration, the three-way partnership between GenBank (NCBI/NLM, US), ENA (EMBL-EBI, UK/EU), and DDBJ (National Institute of Genetics, Japan)...
- rCRS
- The revised Cambridge Reference Sequence (GenBank accession NC_012920), the standard reference for human mitochondrial DNA. All mtDNA variant positions in forensic and...