Regex starts with "ATG" ends with "TAG, TAA orTGA" but does not contain "ATG" and "TAG, TAA or TGA" in between

Question

I'm searching for patterns in a String starting with ATG, ending with TAG, TAA or TGA and length = multiple of 3. ATG and TAG, TAA or TGA can only appear at respectively beginning or end. Which means:

From ATGTTGTGATGT extract ATGTTGTGA

From ATGATGTTGTGATGT extract ATGTTGTGA

Currently I'm using regex (ATG)([ATG]{3})+?(TAG|TAA|TGA).

For ATGATGTTGTGATGT this gets me the wrong result ATGATGTTGTGA. I've tried:

(^ATG)(!?=.*ATG)([ATG]{3})+?(TAG|TAA|TGA)
(^ATG)(!?=(ATG)+)([ATG]{3})+?(TAG|TAA|TGA)

How to tell it to contain ATG only once in the beginning and no more after that?

Your first example does not seems to match your description: it doesn't end with TAG|TAA|TGA? — Ward, Dec 31 '17 at 12:49
This is not possible with regexp. You could do a *negative search* for `ATG` being not first or last and *negate* this result. — Timothy Truckle, Dec 31 '17 at 12:50
Try [`ATG(?:(?!ATG)[ATG]{3})*?(?:TAG|TAA|TGA)`](https://regex101.com/r/GCdGs8/1) — Wiktor Stribiżew, Dec 31 '17 at 12:54
@Ward : I modified the question. Hopefully it's more clear now. — xtra, Dec 31 '17 at 12:59
@WiktorStribiżew : This does exactly what I was looking for in the question — xtra, Dec 31 '17 at 13:00
@xtra Great, I posted as an answer with demos and expanation. — Wiktor Stribiżew, Dec 31 '17 at 13:03
If you're parsing nucleotides, you should probably use a library designed to work with them instead of going for raw regular expressions. — Kayaman, Dec 31 '17 at 13:22

score 2 · Accepted Answer · answered Dec 31 '17 at 13:00

You may use

ATG(?:(?!ATG)[ATG]{3})*?(?:TAG|TAA|TGA)

See the regex demo

Details

ATG - an ATG substring
(?:(?!ATG)[ATG]{3})*? - a tempered greedy token matching any sequence of 3 chars from the [ATG] character set that is not equal to ATG (that is restricted with the negative lookahead (?!ATG))
(?:TAG|TAA|TGA) - either of the three alternatives defined in the non-capturing group: TAG, TAA or TGA.

Java demo:

String rx = "ATG(?:(?!ATG)[ATG]{3})*?(?:TAG|TAA|TGA)";
String s = "ATGTTGTGATGT, ATGATGTTGTGATGT, ATGATGTTGTGATGT";
Pattern pattern = Pattern.compile(rx);
Matcher matcher = pattern.matcher(s);
while (matcher.find()) {
        System.out.println(matcher.group(0));
}

Result:

ATGTTGTGA
ATGTTGTGA
ATGTTGTGA

Regex starts with "ATG" ends with "TAG, TAA orTGA" but does not contain "ATG" and "TAG, TAA or TGA" in between

1 Answers1