Six-Frame Translation and ORF Finder

A double-stranded DNA sequence can be read six ways: three frames on the strand you pasted, shifting the starting point by zero, one and two bases, plus three more on the reverse complement. This page translates all six at once and then marks the ORFs, running from a start codon to the stop codon in the same frame. Positions are always reported on the 1 to N numbering of the strand you supplied, never on the internal numbering of the reverse complement. That is why the negative frames show a start larger than the end: in a 9-base sequence, the ORF in frame -1 runs from position 9 to position 1. The range includes the stop codon, so whenever a stop exists the nucleotide length is three times the amino acid count plus three. The minimum length filter counts amino acids only, leaving the stop out. A start that never meets a stop is still listed, with the stop column saying so.

Three NCBI tables are available. The standard table 1 holds 64 codons and three stops: TAA, TAG and TGA. The vertebrate mitochondrial table 2 changes exactly four of them: AGA and AGG become stops, ATA becomes methionine and TGA becomes tryptophan. The bacterial and plastid table 11 keeps the same amino acid assignments as the standard code; what differs is the set of initiation codons, which includes TTG, CTG, ATT, ATC, ATA, ATG and GTG. Three codons are enough to show it: ATGAGATAA under table 1 yields the peptide MR with a TAA stop ending at position 9, while the same input under table 2 yields only M, because AGA is already a stop at positions 4 to 6. The start selector decides whether ATG alone opens an ORF or whether any initiation codon of the chosen table counts. Feed GTGCCCTAA with table 11 and the ATG mode finds nothing, while the alternative mode returns MP starting at GTG. The initiator is written as M because the ribosome loads it with methionine, and the real codon stays visible in the start codon column.

A few limits worth knowing before you trust a row. IUPAC ambiguity codes are accepted, but any codon containing one translates to X, even where the reading would be certain: CTN is leucine in every combination and still shows up as X. Characters that are not bases are never dropped quietly. The page refuses the input and names the offending character, gap dashes and alignment dots included. FASTA headers, spaces, line breaks and position numbering are stripped, U is read as T and lower case is fine, but only one record at a time: two headers stop the analysis instead of gluing the sequences together. For each stop codon the list reports the ORF beginning at the first start after the previous stop, that is, the longest one in that frame, so internal ATG codons do not become separate rows. Everything else in gene finding is missing: no Shine-Dalgarno search, no codon usage statistics, no intron removal, which means a eukaryotic gene with introns will not give you the mature protein here. A long ORF is not evidence of a gene either, only a stretch without a stop.