SBML.jl — load systems biology models from SBML files

This package provides a straightforward way to load model- and simulation-relevant information from SBML files.

The library provides a single function readSBML to load a Model:

julia> using SBML
julia> mdl = readSBML("Ec_core_flux1.xml")
Model(…)

julia> mdl.compartments
2-element Array{String,1}:
 "Extra_organism"
 "Cytosol"

There are several functions to help you with using the data in the usual COBRA-style workflows, such as getS:

julia> metabolites, reactions, S = getS(mdl)
julia> metabolites
77-element Array{String,1}:
 "M_succoa_c"
 "M_ac_c"
 "M_etoh_c"
  ⋮

julia> S
77×77 SparseArrays.SparseMatrixCSC{Float64,Int64} with 308 stored entries:
  [60,  1]  =  -1.0
  [68,  1]  =  1.0
  [1 ,  2]  =  1.0
  [6 ,  2]  =  -1.0
  ⋮
  [23, 76]  =  1.0
  [56, 76]  =  -1.0
  [30, 77]  =  -1.0
  [48, 77]  =  1.0

julia> Matrix(S)
77×77 Array{Float64,2}:
 0.0   1.0  0.0  0.0  0.0  0.0  0.0  …  0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0   0.0  0.0  0.0  0.0  0.0  0.0     0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0   0.0  0.0  0.0  0.0  0.0  0.0     0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0   0.0  1.0  0.0  0.0  0.0  0.0     0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0   0.0  0.0  0.0  0.0  0.0  0.0     0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0  -1.0  0.0  0.0  0.0  0.0  0.0  …  0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0   0.0  0.0  0.0  0.0  0.0  0.0     0.0   0.0  1.0  -1.0  0.0  0.0  0.0
 0.0   0.0  0.0  0.0  0.0  0.0  0.0     0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0   0.0  0.0  0.0  0.0  0.0  0.0     0.0   0.0  0.0   0.0  0.0  0.0  0.0
 0.0   0.0  0.0  0.0  0.0  0.0  0.0     0.0  -1.0  0.0   0.0  0.0  0.0  0.0
 ⋮                         ⋮         ⋱  ⋮                          ⋮

Functions

Data structures

SBML.Model — Type

Structure that collects the model-related data. Contains units, compartments, species and reactions. The contained dictionaries are indexed by identifiers of the corresponding objects.

source

SBML.Reaction — Type

Reaction with stoichiometry that assigns reactants and products their relative consumption/production rates (accessible in field stoichiometry), lower/upper bounds (in tuples lb and ub, with unit names), and objective coefficient (oc).

source

SBML.Species — Type

Species metadata – contains a human-readable name, and a compartment identifier

source

SBML.UnitPart — Type

Part of a measurement unit definition that corresponds to the SBML definition of Unit. For example, the unit "per square megahour", Mh^(-2), is written as:

UnitPart("second",  # base SI unit, this says we are measuring time
         -2,        # exponent, says "per square"
         6,         # log-10 scale of the unit, says "mega"
         1/3600)    # second-to-hour multiplier

Compound units (such as "volt-amperes" and "dozens of yards per ounce") are built from multiple UnitParts; see the definition of field units in Model.

source

Base functions

SBML.SBMLVersion — Method

function SBMLVersion()

Get the version of the used SBML library in Julia version format.

source

SBML.readSBML — Method

function readSBML(fn::String)::Model

Read the SBML from a XML file in fn and return the contained Model.

source

Data helpers

SBML.getLBs — Method

function getLBs(m::Model)::Vector{Tuple{Float64,String}}

Extract a vector of lower bounds of reaction rates from the model. All bounds are accompanied with the unit of the corresponding value (this behavior is based on SBML specification).

source

SBML.getOCs — Method

function getOCs(m::Model)::Vector{Float64}

Extract the vector of objective coefficients of each reaction.

source

SBML.getS — Method

function getS(m::Model; zeros=spzeros)::Tuple{Vector{String},Vector{String},AbstractMatrix{Float64}}

Extract the vector of species (aka metabolite) identifiers, vector of reaction identifiers, and the (dense) stoichiometry matrix from an existing Model. Returns a tuple with these values.

The matrix is sparse by default (initially constructed by SparseArrays.spzeros). You can fill in a custom empty matrix constructed to argument zeros; e.g. running with zeros=zeros will produce a dense matrix.

source

SBML.getUBs — Method

function getUBs(m::Model)::Vector{Tuple{Float64,String}}

Likewise to getLBs, extract the upper bounds.

source