Tutorials - BioZig

Cross-Domain Interaction PIPELINE DESIGN

In standard bioinformatics, you read a file (Parser), parse it into an object (Domain), and run a method on the object (Algorithm). In BioZig (BZ), data is orthogonal to behavior. The parser generates raw Domain structs, and the Algorithms consume them. No objects own methods.

1. Pure Zig (Native) STRUCTURAL PROTEIN PIPELINE

Parse a PDB file using zero-copy MMap, generate a Structure of Arrays (SoA) coordinate set, and compute its geometric centroid.

const std = @import("std");
const biozig = @import("biozig");
const structural = biozig.algorithms.structural;
const pdb = biozig.ingestion.structural.pdb;

pub fn main() !void {
    var arena = std.heap.ArenaAllocator.init(std.heap.page_allocator);
    defer arena.deinit();
    const allocator = arena.allocator();

    // 1. Zero-copy ingest PDB using 64-bit Bit-Sieve
    const coords = try pdb.parseCoordinateSet(allocator, "protein.pdb");

    // 2. Compute centroid via SIMD vectorized loop
    const centroid = structural.computeCentroid(coords);
    std.debug.print("Centroid: X={d:.3}, Y={d:.3}, Z={d:.3}
", .{centroid.x, centroid.y, centroid.z});
}

2. Python Wrapper SINGLE-CELL TRANSCRIPTOMICS (PCA)

BZ's Python bindings expose C-ABI pointers. We can ingest a massive Matrix Market (.mtx) file and run a fast PCA directly in hardware memory without instantiating millions of Python objects.

import biozig as bz

# 1. Parse sparse single-cell expression data via zero-copy MTX reader
mtx_data = bz.transcriptomics.parse_mtx("cells.mtx")

# 2. Calculate top 10 principal components using the Analytics module
pca_results = bz.analytics.dimensionality.pca(mtx_data, n_components=10)

print(f"Explained Variance: {pca_results.explained_variance}")
print(f"Transform: {pca_results.transform}")

3. R Wrapper SYSTEMS BIOLOGY (COMMUNITY DETECTION)

In R, loading large network graphs often causes RAM exhaustion. BZ handles the graph traversal outside of the R garbage collector.

library(biozig)

# 1. Load systems network (e.g., from an edge list or BioPAX)
graph <- bz_systems_parse_edgelist("protein_network.txt")

# 2. Compute Louvain communities using the Graph builder
communities <- bz_systems_louvain(graph)

print(head(communities))

4. Command Line Interface (CLI) GENOMICS (FASTA METRICS)

The BZ CLI provides direct access to these pipelines from bash. Because the CLI is purely compiled Zig, execution is instantaneous.

# Compute GC Content and Shannon Entropy for a large genome FASTA
$ biozig genomics metrics --input human_genome.fasta --metrics gc_content,entropy

> GC Content: 41.2%
> Shannon Entropy: 1.98 bits