Proteins

A protein opens as a document of its own, with one letter per residue, and its features (domains, sites, mature peptides) drawn on it as they are on DNA.

Opening a protein

• A protein FASTA file (.faa, or any FASTA file). Each record says what it is by its letters: one with letters no base has (E, F, I, L, P, Q, …) opens as a protein. A record written only in nucleotide codes (A, C, G, T, U and the IUPAC ambiguity codes) is always DNA, even a short peptide that happens to be spelt that way, so a DNA file is never taken for a protein.
• A GenPept file (.gp, .gpff): an NCBI protein record, GenBank's format with aa in the LOCUS line. Its header, references and features are read as a GenBank file's are, with feature positions in residues. A site listed as order(...) (the residues a ligand touches, say) stays an order(...) and is written back as one.
• A SnapGene protein file (.prot): its residues, features and notes (see Formats).
• File ▸ Open from NCBI… with a protein accession (NP_000509, AAA12345): the GenPept record is fetched and opens as if you had downloaded it (see Opening a record from NCBI).
• Translate ▸ Open as protein, from a CDS feature or a frame of a DNA document (see below). What the DNA had annotated inside the CDS — a domain, a site, a signal or mature peptide, a misc_feature marking a motif — comes with it, on the residues those bases code for. A feature that runs past the CDS keeps the part inside it and is marked partial there. What is about the DNA itself (a promoter, an intron, a primer site) is left behind.
• New, and choose Protein, to type or paste one.
• Pasting residues with nothing open, without a FASTA header at all: they open as a protein when the text is shaped like a sequence — one block or wrapped lines, of at least ten letters. Nearly every letter is an amino acid, so a sentence would pass the alphabet; the shape is what keeps prose out. Text written only in nucleotide codes is still bases.

What a protein has, and what it does not

The sequence view shows the residues with the ruler counting them, and the toolbar says 147 aa, protein. Features, Find, Style, Case, Undo, History, Compare with… and share links all work as they do for DNA. Find looks for residues as well as feature names: X in what you type matches any residue, and B, Z and J either of the two they stand for.

Everything that reads DNA is gone rather than greyed out: the map and the view switcher, the Complement, Translations and Cut sites toggles and their keys, Reverse complement and Make circular, and the Enzymes, ORFs, Translate, Primers and Cloning tabs, and Detect features, whose parts are DNA. A protein is not offered as a template, insert or vector on the Bench. Typing takes residues only, and residues do not paste into DNA, nor bases into a protein: translate DNA first.

A protein downloads as GenPept (.gp) from File ▸ Download GenPept…, and as protein FASTA (.faa) from Export sequence as FASTA.

Aligning two proteins

The Align tab works in front of a protein: paste or drop another protein — residues, a protein FASTA record or a GenPept file — and it is aligned against the document, globally or locally. Two differences from a DNA alignment:

• It is scored by BLOSUM62, the matrix protein comparisons are usually read against, so a conservative substitution counts in favour rather than against. The match line says which is which: | the same residue, : a substitution the matrix thinks likely (K against R, L against I), . one it does not. Gaps cost what BLAST charges for a protein.
• There is no second strand to try, so nothing is said about one and the alignment is never reported as a reverse complement.

The Protein tab

The Protein tab gives, for the whole protein or the residues selected:

• Length in residues.
• Molecular weight, from average residue masses plus one water.
• Theoretical pI, from the pK values of Bjellqvist et al., with the N-terminus's pK depending on the first residue, as ExPASy ProtParam computes it, and the charge at pH 7.
• The extinction coefficient at 280 nm in water (Pace et al., 1995): 5500 per Trp and 1490 per Tyr, plus 125 per cystine when every pair of Cys forms one. Both values are given, with the absorbance of 1 g/L (Abs 0.1%) for each. A protein with no Trp, Tyr or Cys has none, and says so.
• Composition: how many of each residue, and what share.

These are computed as ExPASy ProtParam computes them, and agree with it. A stop (*) is not counted as a residue. Ambiguous residues (B, Z, X) are weighed by an average and given no charge, and the tab says the numbers are then estimates.

Back-translate

At the foot of the Protein tab, Back-translate writes DNA for the protein, with the same host, codon choice, enzyme sites and GC and run limits as Recode a CDS. A stop (*) is added at the end when the protein has none. It reports the CAI and any limit it could not meet; Open as DNA opens the result as a new document with the CDS annotated, and the DNA is translated again first to check it gives the protein. A protein with ambiguous residues (X, B, Z) has no codon for them and is refused. The protein document itself is not changed.

Open as protein

In a DNA document's Translate tab:

1. Under CDS features, click Open as protein beside a CDS. The protein opens in a new tab named after the CDS, on its Protein tab. It is the CDS's translation as drawn under it in the sequence view: its own genetic code (/transl_table), /codon_start, /transl_except (a selenocysteine shows as U), every piece of a join, and the reverse strand read from its reverse complement, and the residue two bases at a partial 3′ end give when they fix it (see Translation). The stop that ends it is left off; a stop inside it stays, as *, since it is what the bases say.
2. Or select a stretch of DNA and click Open as protein on one of the six frames. That frame's translation opens as it stands, stops and all, named after the document, the range and the frame, and read with the genetic code chosen in the tab.

The protein carries the DNA's organism, and a comment saying what it was translated from. The DNA is not changed.