the reading frame
Read this string of letters with no spaces: THEFATCATATETHERAT. Start at the first letter and group by threes: THE FAT CAT ATE THE RAT — perfectly sensible. Now start one letter late: H EFA TCA TAT ETH ERA T — nonsense. The letters never changed; only where you began did. That choice of where to start counting threes is the reading frame.
Because the genetic code is read three bases at a time with no punctuation between codons, any stretch of RNA can be read in three different frames depending on where the first codon begins. The cell does not guess; the start codon AUG sets the frame, and from that point every group of three is locked in until a stop codon is reached. A continuous run of codons with a start and an in-frame stop, long enough to encode a protein, is called an open reading frame — the signature a gene-finding computer program hunts for in a freshly sequenced genome.
The reading frame is fragile in a specific, dramatic way. If a single base is inserted or deleted within a gene, every codon downstream shifts by one and the message turns to gibberish — a frameshift mutation. This is why an indel (insertion or deletion) of one or two bases is usually far more damaging than swapping one base for another, and why inserting or deleting a multiple of three bases is comparatively gentle: it adds or removes whole codons without throwing off the frame. The reading frame is the silent assumption behind every protein the cell makes.
The mRNA 5'-AUGCAUGGU-3' in its proper frame reads AUG CAU GGU = Met-His-Gly. Delete the first C and it becomes 5'-AUGAUGGU-3', read AUG AUG GU = Met-Met-... — a completely different protein from the same letters.
Deleting one base shifts the frame, scrambling every codon that follows.
A long open reading frame strongly hints at a real protein-coding gene, but it is not proof — the cell still has to actually transcribe and translate it. Computers find candidate ORFs; biology confirms which ones are truly used.