All on iteration

When working with molecular entities, we want to run over all atoms of a system, over all chains, etc. In this tutorial we will learn how this can be done.

Molecular systems

In BiochemicalAlgorithms.jl atoms and bonds are existing inside a System. Typically, molecular data is stored in molecular data formats such as PDB. The latter can be directly read into a system.

sys = load_pdb(ball_data_path("../test/data/AlaAla.pdb"))

System{Float32} with 23 atoms (AlaAla.pdb)

You can, e.g., print all atoms of the given system as a table:

atoms(sys)

AtomTable{Float32} with 23 rows:

Single columns can be directly accessed by their name:

atoms(sys).name

23-element SystemComponentTableCol{String}:
 "N"
 "CA"
 "C"
 "O"
 "1H"
 "1HB"
 "2H"
 "2HB"
 "3H"
 "3HB"
 ⋮
 "C"
 "H"
 "1HB"
 "2HB"
 "3HB"
 "CB"
 "HA"
 "O"
 "OXT"

You can also directly query the number of atoms via natoms:

natoms(sys)

23

Similar functions exist for other system components, including bonds, molecules, chains, and fragments.

How can I iterate over all atoms of a system?

We can just as easily iterate over all atoms:

for atom in atoms(sys)
    # do something with this atom
end

Or, for individual columns:

for atom_name in atoms(sys).name
    # do something with this atom name
end

How can I iterate over specific atoms?

In many scenarios, we only want to iterate over a subset of atoms fulfilling a specific criteria. For example, here we only want to get the positions of the $C_\alpha$ atoms:

ca_atoms = filter(atom -> atom.name == "CA", atoms(sys))

AtomTable{Float32} with 2 rows:

And here we only want the heavy atoms:

heavy_atoms = filter(atom -> atom.element != Elements.H, atoms(sys))

AtomTable{Float32} with 11 rows:

Instead of a system, we can use any other system component as an argument to the atoms function to only list the atoms of the same component. For example, the following system contains two chains:

sys = load_pdb(ball_data_path("../test/data/2ptc.pdb"))
chains(sys)

ChainTable{Float32} with 2 rows:

In order to get all atoms of the first chain, we can use:

chainE = first(chains(sys))
for atom in atoms(chainE)
    # do something with this atom of the first chain
end

We can even combine this approach with filtering to get, e.g., only the $C_\alpha$ atoms of the first chain:

for atom in filter(atom -> atom.name == "CA", atoms(chainE))
    # do something with this CA atom of the first chain
end

How can I iterate over all bonds of a system?

Bonds are not explicitely stored in the PDB format but are rather inferred after reading the data into a system using the fragment database FragmentDB:

sys = load_pdb(ball_data_path("../test/data/AlaAla.pdb"))

# bonds are not contained in the pdb file
nbonds(sys)

0

# use the fragment database for normalizing naming schemas between molecular file formats, reconstruction of missing parts of the structure and building the bonds
fdb = FragmentDB()

normalize_names!(sys, fdb)
reconstruct_fragments!(sys, fdb)
build_bonds!(sys, fdb)

nbonds(sys)

[ Info: reconstruct_fragments!(): added 0 atoms.
[ Info: build_bonds!(): built 22 bonds

22

Similar to the atom iteration, we can iterate over all bonds of a sysem:

for bond in bonds(sys)
    # do something with this bond
end

How can I iterate over all fragments/residues/nucleotides of a system?

Chains usually contain fragments, e. g., residues or nucleotides and can either be queried collectively…

sys = load_pdb(ball_data_path("../test/data/2ptc.pdb"))

nfragments(sys)

439

…or by a given type:

println("Number of residues: ", nresidues(sys))
println("Number of nucleotides: ", nnucleotides(sys))

Number of residues: 281
Number of nucleotides: 0

Just like in the examples above, the functions can also be used with other atom containers:

nresidues.(chains(sys))

2-element Vector{Int64}:
 223
  58