Translation

Translating DNA to AA¶

You can translate DNA FASTA to amino acids in all six reading frames using dna-to-aa.py. Start with this FASTA file:

$ cat dna.fasta
>SK7F6:25:89
GTTACAATGAATG
>SK7F6:30:80
TCAGCCTCACCGCTGTGTCAAATGTACACAGATACTTTTCCGGCTCCAGCGCTGGCATGAAAACAAAATTAACAATTATTTTTTGA

On translating, you'll see what you'd expect:

$ dna-to-aa.py < dna.fasta
>SK7F6:25:89-frame0
VTMNX
>SK7F6:25:89-frame1
LQ*M
>SK7F6:25:89-frame2
YNEX
>SK7F6:25:89-frame0rc
HSL*X
>SK7F6:25:89-frame1rc
IHCN
>SK7F6:25:89-frame2rc
FIVT
>SK7F6:30:80-frame0
SASPLCQMYTDTFPAPALA*KQN*QLFFX
>SK7F6:30:80-frame1
QPHRCVKCTQILFRLQRWHENKINNYFLX
>SK7F6:30:80-frame2
SLTAVSNVHRYFSGSSAGMKTKLTIIF*
>SK7F6:30:80-frame0rc
SKNNC*FCFHASAGAGKVSVYI*HSGEAX
>SK7F6:30:80-frame1rc
QKIIVNFVFMPALEPEKYLCTFDTAVRLX
>SK7F6:30:80-frame2rc
KK*LLILFSCQRWSRKSICVHLTQR*G*

There are 6 translations for each sequence.

You can see that dna-to-aa.py adds frame information to the read ids (rc = reverse complement).

Any start (M) and stop (*) codons in the translations are left intact.

Translating RNA¶

If your reads are RNA, you can pass the type using --type rna:

$ dna-to-aa.py --type rna < dna-reads.fasta

Requiring a sufficiently long ORF¶

To only output translations that contain at least one ORF of at least some specified length:

$ dna-to-aa.py --minORFLength 20 < dna.fasta
>SK7F6:30:80-frame1
QPHRCVKCTQILFRLQRWHENKINNYFLX
>SK7F6:30:80-frame2
SLTAVSNVHRYFSGSSAGMKTKLTIIF*
>SK7F6:30:80-frame1rc
QKIIVNFVFMPALEPEKYLCTFDTAVRLX

Extracting ORFs¶

To find and extract all ORFs in the translations, use extract-ORFs.py:

$ extract-ORFs.py < dna.fasta
>SK7F6:25:89-frame0-[0:5)
VTMNX
>SK7F6:25:89-frame1-(0:2]
LQ
>SK7F6:25:89-frame2-(0:4)
YNEX
>SK7F6:25:89-frame0rc-(0:3]
HSL
>SK7F6:25:89-frame1rc-(0:4)
IHCN
>SK7F6:25:89-frame2rc-(0:4)
FIVT
>SK7F6:30:80-frame0-[0:19]
SASPLCQMYTDTFPAPALA
>SK7F6:30:80-frame1-(0:29)
QPHRCVKCTQILFRLQRWHENKINNYFLX
>SK7F6:30:80-frame2-[0:27]
SLTAVSNVHRYFSGSSAGMKTKLTIIF
>SK7F6:30:80-frame0rc-(0:5]
SKNNC
>SK7F6:30:80-frame1rc-[0:29)
QKIIVNFVFMPALEPEKYLCTFDTAVRLX
>SK7F6:30:80-frame2rc-(0:2]
KK

This can also be restricted by minimal ORF length:

$ extract-ORFs.py --minORFLength 20 < dna.fasta
>SK7F6:30:80-frame1-(0:29)
QPHRCVKCTQILFRLQRWHENKINNYFLX
>SK7F6:30:80-frame2-[0:27]
SLTAVSNVHRYFSGSSAGMKTKLTIIF
>SK7F6:30:80-frame1rc-[0:29)
QKIIVNFVFMPALEPEKYLCTFDTAVRLX

Unlike dna-to-aa.py, extract-ORFs.py writes one new sequence per ORF and start and stop colons do not appear in the output.

extract-ORFs.py also adds a suffix to the read ids to indicate the ORF offset in the original sequence, e.g., the [0:29) in SK7F6:30:80-frame1rc-[0:29). A parenthesis indicates that that side of the ORF was open (see below), whereas a square bracket indicates that the ORF had a start or stop codon on that side. The numbers in between give the start and stop offsets using Python's convention that the stop offset is not part of the ORF.

To illustrate, the id SK7F6:30:80-frame1rc-[0:29) tells us that the original read was SK7F6:30:80 (TCAGCCTCACCG...ATTATTTTTTGA). That was reverse complemented (rc) to get TCAAAAAATAAT...CGGTGAGGCTGA. The translation was of the frame with offset 1: CAAAAAATAAT...CGGTGAGGCTGA, producing `. In that translation, the ORF had Python offsets (0, 29), corresponding to the AA substringQKIIVNFVFMPALEPEKYLCTFDTAVRLX` (as shown above).

Include "open" ORFs¶

Some reads start or end in the middle of an ORF. In this case, the ORF is considered "open" on either the left or right (or both). It is not possible to tell how long these ORFs actually are, we just have a minimum. If you'd like these ORFs to be printed even though we do not know if they meet the length requirement, use --allowOpenORFs:

$ extract-ORFs.py  --minORFLength 20 --allowOpenORFs < dna.fasta
>SK7F6:25:89-frame0-[0:5)
VTMNX
>SK7F6:25:89-frame1-(0:2]
LQ
>SK7F6:25:89-frame2-(0:4)
YNEX
>SK7F6:25:89-frame0rc-(0:3]
HSL
>SK7F6:25:89-frame1rc-(0:4)
IHCN
>SK7F6:25:89-frame2rc-(0:4)
FIVT
>SK7F6:30:80-frame1-(0:29)
QPHRCVKCTQILFRLQRWHENKINNYFLX
>SK7F6:30:80-frame2-[0:27]
SLTAVSNVHRYFSGSSAGMKTKLTIIF
>SK7F6:30:80-frame0rc-(0:5]
SKNNC
>SK7F6:30:80-frame1rc-[0:29)
QKIIVNFVFMPALEPEKYLCTFDTAVRLX
>SK7F6:30:80-frame2rc-(0:2]
KK

Starting from AA input¶

As well as accepting DNA FASTA, extract-ORFs.py can also handle AA FASTA input, but you have to tell it. So if you already have a file of AA FASTA and you want the ORFs, use extract-ORFs.py --type aa.

You'll probably never need to do this (because you can use extract-ORFs.py to do the complete job), but you can convert DNA FASTA to AA using dna-to-aa.py and pipe that into extract-ORFs.py:

$ dna-to-aa.py < dna.fasta | extract-ORFs.py --minORFLength 20 --type aa
>SK7F6:30:80-frame1-(0:29)
QPHRCVKCTQILFRLQRWHENKINNYFLX
>SK7F6:30:80-frame2-[0:27]
SLTAVSNVHRYFSGSSAGMKTKLTIIF
>SK7F6:30:80-frame1rc-[0:29)
QKIIVNFVFMPALEPEKYLCTFDTAVRLX

In this case, if you forget to tell extract-ORFs.py that its input is AA, you'll get a translation error and a reminder to use --type aa:

$ dna-to-aa.py < dna.fasta | extract-ORFs.py
Could not translate read 'SK7F6:25:89-frame0' sequence 'VTMNX' (Codon 'NXN' is invalid).
Did you forget to run extract-ORFs.py with "--type aa"?