Getting Started

The typical SHEPHARD workflow has three steps: (1) build a Proteome from a FASTA file, (2) layer on annotations from SHEPHARD files (or compute them on the fly), and (3) query the relationship between sequence and those annotations.

A minimal end-to-end example

The following code reads a UniProt FASTA file, annotates it with domains, and then prints every protein that has a domain starting at its immediate N-terminus.

from shephard.apis import uniprot
from shephard.interfaces import si_domains

# read a UniProt-formatted FASTA file into a new Proteome
P = uniprot.uniprot_fasta_to_proteome('fasta_file.fasta')

# annotate every protein with domains from a SHEPHARD domains file
si_domains.add_domains_from_file(P, 'domains_file.tsv')

# iterate the Proteome -> Protein -> Domain hierarchy
for protein in P:
    for domain in protein.domains:
        if domain.start == 1:
            print(f'Protein {protein} has an N-terminal domain: {domain}')

A few things to note:

• Iterating a Proteome (for protein in P) yields Protein objects directly.

• protein.domains returns a list of Domain objects (sorted by start position) — you iterate the objects directly; you do not index back through the protein.

• All positions use biology-style indexing (1-based, inclusive), so domain.start == 1 really is the first residue.

Building a Proteome by hand

You do not need a file to get started — Proteins can be added programmatically:

from shephard import Proteome

P = Proteome()
P.add_protein('MAEPQRDGSALLK', 'my cool protein', 'test1')

prot = P.protein('test1')
prot.add_domain(1, 5, 'IDR')
prot.add_site(4, 'phosphorylation', symbol='P', value=1.0)

print(prot.get_sequence_region(1, 5))      # 'MAEPQ'
print(prot.get_domains_by_type('IDR'))     # [ |Domain: IDR ... ]
print(prot.site(4))                        # [ |Site: phosphorylation ... ]

For more worked examples, see Examples.