Get Started

Installation

To install Julia, please go to the official Julia website. Please see platform specific instructions if you have trouble installing Julia.

Requirements: Julia 1.9 or later.

To install the NNMM package, use the following command inside the Julia REPL (or IJulia Notebook):

using Pkg
Pkg.add("NNMM")

To load the NNMM package:

using NNMM

Development Version

To use the latest/beta features under development:

Pkg.add(url="https://github.com/reworkhow/NNMM.jl")

Jupyter Notebook

If you prefer "reproducible research", an interactive Jupyter Notebook interface is available for Julia (and therefore NNMM). The Jupyter Notebook is an open-source web application for creating and sharing documents that contain live code, equations, visualizations and explanatory text. To install IJulia for Jupyter Notebook, please go to IJulia.

Multi-threaded Parallelism

NNMM supports multi-threaded parallelism for faster computation. The number of threads can be checked by:

Threads.nthreads()

To start Julia with multiple threads (requires Julia 1.5+):

julia --threads 4

Or set the environment variable:

export JULIA_NUM_THREADS=4
julia

Access Documentation

To show basic information about NNMM in REPL or IJulia notebook, use ?NNMM and press enter.

For help on a specific function, type ? followed by its name:

?runNNMM   # Help on runNNMM function
?Layer     # Help on Layer type
?Equation  # Help on Equation type

The full documentation is available here.

Quick Example

using NNMM
using NNMM.Datasets
using DataFrames, CSV
using Random

# Set seed for reproducibility
Random.seed!(42)

# Load built-in example data
geno_path = Datasets.dataset("genotypes_1000snps.txt", dataset_name="simulated_omics_data")
pheno_path = Datasets.dataset("phenotypes_sim.txt", dataset_name="simulated_omics_data")
pheno_df = CSV.read(pheno_path, DataFrame)

# Create omics and trait files
omics_df = pheno_df[:, vcat(:ID, [Symbol("omic$i") for i in 1:10])]
CSV.write("omics.csv", omics_df; missingstring="NA")
CSV.write("trait.csv", pheno_df[:, [:ID, :trait1]]; missingstring="NA")

# Define model architecture
layers = [
    Layer(layer_name="geno", data_path=[geno_path]),
    Layer(layer_name="omics", data_path="omics.csv", missing_value="NA"),
    Layer(layer_name="pheno", data_path="trait.csv", missing_value="NA")
]

equations = [
    Equation(
        from_layer_name="geno", to_layer_name="omics",
        equation="omics = intercept + geno",
        traits=["omic$i" for i in 1:10],
        method="BayesC"
    ),
    Equation(
        from_layer_name="omics", to_layer_name="pheno",
        equation="pheno = intercept + omics",
        traits=["trait1"],
        activation_function="linear"
    )
]

# Run MCMC analysis
results = runNNMM(layers, equations;
    chain_length=5000,
    burnin=1000,
    output_folder="my_results"
)

# Access EBV results
ebv = results["EBV_NonLinear"]
println("Mean EBV: ", mean(ebv.EBV))

# Cleanup
rm("omics.csv", force=true)
rm("trait.csv", force=true)

Run Your Analysis

There are several ways to run your analysis:

Interactive Session (REPL)

Start an interactive session by double-clicking the Julia executable or running julia from the command line:

julia> using NNMM
julia> # your analysis code here

To evaluate code written in a file script.jl in REPL:

julia> include("script.jl")

To exit the interactive session, type ^D (control + d) or quit().

Command Line

To run code in a file non-interactively from the command line:

julia script.jl

If you want to pass arguments to your script:

julia script.jl arg1 arg2

where arguments arg1 and arg2 are passed as ARGS[1] and ARGS[2] of type String.

Jupyter Notebook

To run code in Jupyter Notebook, please see IJulia.

Key Concepts

Three-Layer Architecture

NNMM uses a three-layer neural network:

1. Layer 1 (Input): Genotypes (SNP markers)
2. Layer 2 (Hidden): Omics/Latent traits (can have missing values)
3. Layer 3 (Output): Phenotypes

Core Types

Core Function

Next Steps

• Tutorial: Complete step-by-step analysis
• Part 2: Basic NNMM: Fully-connected networks
• Part 3: NNMM with Omics: Using observed omics data
• API Reference: Full function documentation