Protease Digest

1. 01Forty-nine proteases and cleavage chemicals, from trypsin, chymotrypsin, Lys-C, Asp-N and CNBr to the ten caspases, Factor Xa, enterokinase, thrombin and pepsin at two pH values
2. 02TEV protease, sequencing-grade thrombin, ProAlanase, papain, ficin, bromelain, elastase and alpha-lytic protease, so a fusion tag can be cut in silico before it is cut on the bench
3. 03One enzyme or several, in two modes that are different experiments: sequential runs each protease on the intact protein, concurrent puts them all in one tube and re-digests the products until nothing changes
4. 04The specificity rule behind every enzyme, shown in the panel with the source it came from, so a result can be cross-checked against the tool it is being compared with
5. 05Missed cleavages to any depth, reported per peptide so a partially cleaved fragment is distinguishable from a fully cleaved one
6. 06Semi-enzymatic digestion, with the terminus that is not a real cleavage site marked as such
7. 07Keil's trypsin exception rules as a toggle, live only for trypsin because that is whose rules they are
8. 08Average and monoisotopic mass per peptide, with m/z for any of the first five protonation states — 1+ through 5+, chosen by you rather than fixed, with 2+ and 3+ on by default because that is where tryptic peptides usually land in electrospray
9. 09Peptide length window and a monoisotopic mass window, both narrowing the list the way an MS inclusion list is narrowed
10. 10Sequence coverage of the peptides actually shown, so a filter that discards half the protein says so
11. 11Cleavage-site count for the whole protein alongside the filtered peptide list
12. 12Reads the live editor selection or a chosen coding feature, translating it in the genetic code you pick
13. 13Any peptide selects its own residues back on the map, strand-aware, so a fragment found in a spectrum can be located on the construct
14. 14CSV export carrying every peptide, its masses and an m/z column per charge state you have switched on, with the run settings and the sample preparation in a header block so the export is reproducible

The digest ports pyteomics `parser.icleave` faithfully, including its ring buffer over cleavage sites and its end-of-sequence sentinel, so missed cleavages and semi-enzymatic peptides come out in the same order and multiplicity as upstream. The rule set merges pyteomics with Bioconductor cleaver, which contributes five enzymes pyteomics lacks, and RPG, whose decision-tree rules contribute ten more that neither defines; the multi-enzyme modes are ported from RPG as well. Every rule and both algorithms are golden-validated against the real libraries. The panel runs on @ports/protease.

Everything runs client-side. Files are parsed in the browser and projects are saved on this machine, so an unpublished sequence is never uploaded.

Which proteases are included?

Forty-nine. Everything ExPASy PeptideCutter offers: trypsin in three specificities, chymotrypsin in two, Arg-C, Asp-N, Lys-C, Lys-N, LysargiNase, glutamyl endopeptidase, proteinase K, thermolysin, pepsin at pH 1.3 and above pH 2, neutrophil elastase, proline endopeptidase, staphylococcal peptidase I, the ten caspases, granzyme B, thrombin, Factor Xa, enterokinase, clostripain, and the chemical cleavages CNBr, BNPS-skatole, iodosobenzoic acid, hydroxylamine, formic acid and NTCB. Plus ten that PeptideCutter does not define, taken from RPG with the sources it cites: TEV protease, sequencing-grade thrombin, sequencing-grade Asp-N, Glu-C in its D-and-E specificity, ProAlanase, papain, ficin, bromelain, elastase and alpha-lytic protease.

What is the difference between Glu-C and glutamyl endopeptidase here, or between the two Asp-N entries?

Buffer, and it changes the peptide list. Glutamyl endopeptidase is the specificity PeptideCutter publishes, cutting after E only, as V8 protease behaves in ammonium bicarbonate; the Glu-C entry cuts after D and E, as it behaves in phosphate. The same applies to Asp-N: the endopeptidase entry cuts before D, while the sequencing-grade entry also cuts before C. Both members of each pair are shipped because they are different reagents, not different opinions about one.

What does concurrent digestion do that running the enzymes one at a time does not?

It can cut where neither enzyme cuts the intact protein. Most cleavage rules inspect the residues around the site, so once one protease has cut, a fragment now ends where that context used to be and a blocked site can open; the products are re-digested until nothing changes. It can also lose a site, when the removed context was required. The order of the enzymes is therefore part of the experiment, and the panel numbers them for that reason.

Three of the cleavage rules differ from pyteomics. Why?

Because ExPASy is the authority for these rules and pyteomics disagrees with it in three places. Its thermolysin rule adds a P2-prime constraint the published table does not contain, and its two pepsin rules block a site near the N-terminus merely because the residues before it do not exist, when the constraint is satisfied vacuously there. Both the adopted rule and the upstream one it replaces are pinned in the engine, so neither a silent revert nor a further drift can pass unnoticed.

Are the masses monoisotopic or average?

Both, per peptide. The mass window filters on the monoisotopic value, since that is what an inclusion list is usually built on, and the CSV carries both alongside an m/z column for every charge state you have switched on — 1+ through 5+, with 2+ and 3+ on by default.

Why is the coverage lower than I expect?

Coverage is computed over the peptides shown, not over every peptide the digest produced. A length or mass window that discards short fragments lowers it, which is the point: it states how much of the protein the filtered list can actually account for.

Can it digest a coding sequence rather than a protein?

Yes. Point it at a selection or a CDS feature and it translates in the genetic code you choose before digesting, and any peptide in the result selects the DNA that encodes it back on the map.

Does it need the protein pasted in?

No. It reads whatever is open: the current selection, a coding feature picked from the construct, or pasted residues. A target it cannot use says why before the run rather than returning an empty table.