An open reading frame (ORF) is a stretch of DNA or RNA that can be read as consecutive three-nucleotide codons without encountering an internal termination codon in the specified reading frame. In a common operational definition, an ORF begins with a potential initiation codon and ends at an in-frame termination codon. Other definitions describe any uninterrupted stretch between stop codons, without requiring an identified start codon. ORFs are used to identify candidate protein-coding sequences, but their presence alone does not establish that a sequence is expressed, translated, or biologically functional.
Reading frames and codon boundaries
During translation, a ribosome reads a messenger RNA in the -to- direction, interpreting successive codons according to the genetic code. Each codon normally specifies an amino acid or a signal to terminate translation. Because codons contain three nucleotides, a single RNA sequence has three possible reading frames, depending on whether grouping begins at the first, second, or third nucleotide.
Double-stranded DNA has six potential reading frames: three on each strand, with each strand read in its own -to- direction. Gene-finding programs commonly inspect all six frames when the transcribed strand is unknown. This does not mean that all six frames are translated.
Under the standard genetic code, the stop codons are UAA, UAG, and UGA in RNA, conventionally written TAA, TAG, and TGA in DNA sequence analysis. AUG is the most common initiation codon, corresponding to ATG in DNA. Alternative initiation codons occur, particularly in bacteria and archaea. The codon table and permitted initiation codons must therefore be appropriate to the organism and cellular compartment being studied.
For example, the DNA sequence
5′-ATG GAA TTT TGA-3′
contains a potential ORF beginning at ATG and ending at TGA. Its corresponding RNA sequence could encode methionine–glutamate–phenylalanine before termination. The stop codon does not itself encode an amino acid.
Definitions and boundaries
“Open reading frame” is used with several related conventions:
- Start-to-stop ORF: a sequence beginning with an accepted initiation codon and continuing to the first in-frame stop codon.
- Stop-to-stop ORF: an uninterrupted interval between in-frame stop codons, within which one or more potential initiation sites may occur.
- Partial ORF: a sequence lacking an identified beginning or end because the available sequence is incomplete or extends beyond the analyzed region.
Software packages differ in which convention they use, whether terminal stop codons are included in reported lengths, and whether partial ORFs are returned. A reported ORF length or coordinate range is therefore meaningful only in conjunction with those settings.
An ORF is also distinct from a coding sequence or CDS. ORF usually describes a sequence feature or a prediction; CDS denotes the region annotated as encoding a particular protein product. A CDS may itself be computationally predicted rather than experimentally established, so this distinction does not provide an automatic hierarchy of experimental certainty.
Neither term is interchangeable with gene. A protein-coding gene can contain untranslated regions and intervening sequences in addition to its coding region, while many genes produce functional RNAs rather than proteins.
ORFs in different biological contexts
In many bacterial and archaeal genes, the coding region is continuous in genomic DNA, making ORF detection a useful initial step in gene prediction. One messenger RNA may contain several protein-coding ORFs, as occurs in polycistronic transcripts.
In eukaryotes, protein-coding regions are often interrupted by introns. RNA splicing joins exons to produce a mature transcript containing a continuous coding sequence. Searching genomic DNA for long uninterrupted ORFs can therefore miss genes whose coding regions are divided among multiple exons. Alternative splicing can generate transcripts with different coding boundaries or protein products from the same gene.
Overlapping ORFs occupy some of the same nucleotides but use different reading frames or initiation sites. They are especially prominent in compact viral genomes, although they also occur in cellular organisms. Their existence means that a single nucleotide change can affect more than one encoded product.
A short open reading frame, or sORF, is an ORF encoding a relatively short peptide. There is no universally applied length cutoff. Some sORFs encode functional peptides, whereas others occur by chance or are translated without an established peptide function.
An upstream open reading frame, or uORF, lies in the leader of a transcript, upstream of its main coding sequence. Translation of a uORF can influence translation of the downstream ORF through mechanisms involving ribosome scanning, termination, and reinitiation. Its regulatory effect depends on the transcript and cellular conditions.
Detection and evidence for translation
Basic ORF detection groups a sequence into codons in each relevant frame, identifies permitted initiation and termination codons, and reports intervals satisfying specified length and boundary criteria. It is a sequence-analysis operation rather than a direct measurement of gene expression.
More complete gene-prediction methods combine ORF structure with evidence such as:
- organism-specific codon frequencies and sequence composition;
- plausible translation-initiation signals;
- splice sites and transcript structures;
- similarity to established proteins;
- conservation across related organisms;
- experimental evidence of transcription or translation.
RNA sequencing can establish that a region is transcribed and help reconstruct transcript boundaries, but transcription alone does not establish protein production. Ribosome profiling identifies RNA fragments protected by ribosomes and can reveal translation in a particular frame. Interpreting these data requires attention to experimental artifacts and the distinction between productive translation and other ribosome-associated events.
Detection of matching peptides by mass spectrometry supplies additional evidence that a predicted product is made. Small, unstable, low-abundance, or condition-specific products may nevertheless escape detection. Evidence that an ORF is translated must also be distinguished from evidence that its product has a biological function.
Applications and limitations
ORF analysis is central to genome annotation, analysis of newly sequenced organisms, and characterization of viral coding capacity. In recombinant DNA work, checking an ORF helps determine whether a proposed construct preserves the intended reading frame and avoids premature termination codons.
Long ORFs are useful clues to protein-coding potential because stop-free stretches become less likely by chance as their length increases. However, their significance depends strongly on nucleotide composition. GC-rich sequences, for example, can contain long accidental ORFs because all three standard stop codons contain relatively high proportions of A and T. Conversely, genuine short coding sequences may be difficult to distinguish from chance ORFs.
A mutation that introduces a premature stop codon can shorten an ORF. An insertion or deletion whose length is not divisible by three can shift the reading frame, altering downstream codons and often introducing a new stop codon. Sequencing and assembly errors can imitate these changes and cause gene-prediction errors.
Standard ORF searches also have biological exceptions. Some organisms and organelles use genetic codes in which particular codons have different meanings. Certain genes employ programmed ribosomal frameshifting or stop-codon readthrough, and specialized mechanisms allow codons normally interpreted as termination signals to specify selenocysteine or pyrrolysine. In these cases, a protein-coding sequence cannot always be described accurately as a single uninterrupted ORF under the standard code.