SeqBench

BLAST Search — Identify a Sequence Against GenBank

Sequence identification against GenBank, with automatic program selection.

Run an NCBI BLAST search from the browser to identify a sequence of unknown provenance: paste a colony-PCR read, an insert or a protein, and it is searched against the whole of GenBank through NCBI's own BLAST service. The program is chosen from the sequence itself: nucleotide goes to megablast against the nucleotide collection (or plain blastn for a short query such as a primer), protein to blastp against the non-redundant protein database. Results come back as a ranked table with each hit's accession, description, organism, percent identity, expect value and how much of YOUR query it covers, with overlapping alignments merged, so coverage cannot read above 100%. A search usually takes twenty to sixty seconds; NCBI asks that a single search is checked no more than once a minute, so the page waits between checks and shows you the request id, which stays valid on their side for about a day.

Paste a sequence to search it against GenBank.

Design, edit, and verify complete constructsMaps, primers, cloning, CRISPR, Sanger, and GenBank/SnapGene files in one workspace.

How to use the BLAST Search tool

  1. 1Paste a DNA or protein sequence, raw or as FASTA.
  2. 2Press Search; the program and database are chosen from the sequence.
  3. 3Read the ranked hits: identity and query coverage together tell you whether a hit explains your whole sequence or only part of it.

FAQ

What is the difference between percent identity and query coverage?

Identity is how well the matched region matches; coverage is how much of your sequence was matched at all. A hit can be high in one and low in the other: a 100% identity hit covering 12% of your query means a short stretch matched perfectly and the remainder of your sequence is unaccounted for, the usual signature of a shared promoter or tag rather than the gene you were looking for. Both are shown side by side.

Why does a search take so long?

The search runs on NCBI's machines against a database of well over a hundred million sequences, queued alongside everyone else's. Twenty to sixty seconds is normal for a nucleotide search; a protein search can take several minutes. NCBI also asks that any one search is polled no more than once a minute, so this page waits between checks. The request id shown while it runs stays valid on their side for about twenty-four hours.

Why does one hit say a number of identical entries?

BLAST collapses database records with the same sequence into a single hit. A sequence deposited a dozen times under different accessions returns one row and a count of the others. The accession shown is the representative one.

Can I choose the program or database myself?

No. The program is inferred from the sequence, because it is not really a choice: a protein query cannot be run through blastn, and a mismatched pair is refused with an explanation rather than quietly returning nothing.

Helpful?