๏‚›GitHub๏„

Quick Start Guide for Juscan.jl

This guide walks you through a basic single-cell RNA-seq analysis pipeline using Juscan.jl. The dataset used in this example comes from doi:10.6084/m9.figshare.22716739.v1.

๐Ÿ“ฆ Prerequisites

Make sure you have installed Juscan.jl, Muon.jl, and their dependencies:

using Pkg
Pkg.activate(".")
Pkg.add(url="https://github.com/zehua0417/Juscan.jl")

๐Ÿ“ Load Data

using Juscan
using Muon
using DataFrames

adata = Juscan.readh5ad("data/data.h5ad")

๐Ÿ” Quality Control

Juscan.Pp.filter_cells!(adata, min_genes=100)
Juscan.Pp.filter_genes!(adata, min_cells=3)
Juscan.Pp.filter_cells!(adata, max_genes=8000)
Juscan.Pp.filter_cells!(adata, max_counts=140000)

adata.var.mt = startswith.(adata.var_names, "MT-")
adata.var.ribo = startswith.(adata.var_names, "RPS") .| startswith.(adata.var_names, "RPL")
adata.var.hb = occursin.(r"^HB[^P]", adata.var_names)

Juscan.Pp.calculate_qc_metrics!(adata, qc_vars=["mt", "ribo", "hb"])

๐Ÿ“Š Visualize QC Metrics

Juscan.Pl.violin(
  adata,
  ["pct_counts_mt", "n_genes_by_counts", "total_counts"];
  width=300,
  height=800,
  fill_alpha=0.7,
  savefig="/home/lihuax/Pictures/Juscan/qc_violin.png",
)

Juscan.Pl.scatter(
  adata,
  "total_counts",
  "n_genes_by_counts",
  color_key="pct_counts_mt",
  width=800,
  height=800,
  colormap_name="magma",
  savefig="/home/lihuax/Pictures/Juscan/qc_scatter.png",
)

๐Ÿ”ฌ Normalization

adata.layers["normalized"] = deepcopy(adata.X)
Juscan.Pp.normalize_total!(adata, target_sum=1000, layer="normalized")
Juscan.Tl.logp1_transform!(adata, layer="normalized", key_added="normalized_logp1")
adata.layers["normalized_logp1"] = Float64.(adata.layers["normalized_logp1"])

๐Ÿงฌ Highly Variable Genes

Juscan.Tl.highly_variable_genes!(adata, n_top_genes=2000, layer="normalized_logp1")
Juscan.Pl.hvg_scatter(adata, savefig="/home/lihuax/Pictures/Juscan/hvg_scatter.png")

๐Ÿ“‰ Dimensionality Reduction

Juscan.Tl.pca!(adata; key_added="pca", n_pcs=15)
Juscan.Tl.pca!(adata; layer="normalized_logp1", key_added="pca", n_pcs=15)
Juscan.Pl.plot_variance_ratio(adata, savefig="/home/lihuax/Pictures/Juscan/variance_ratio.png")

Juscan.Tl.subset_to_hvg!(adata; layer="normalized_logp1", n_top_genes=2000)
Juscan.Pp.filter_cells!(adata, min_counts=3000)
Juscan.Pp.filter_cells!(adata, max_counts=10000)

๐Ÿ”— Clustering & UMAP

Juscan.Tl.clustering!(adata, method="km", use_pca=15, cluster_K=4, dist="Euclidean")
Juscan.Tl.umap!(
  adata;
  key_added="umap",
  use_pca="pca",
  n_pcs=15,
  min_dist=0.5,
  n_neighbors=50,
)
fig = Juscan.Pl.plot_umap(adata, color_by="clusters_0.5")

This pipeline demonstrates the essential steps of scRNA-seq analysis using Juscan.jl. For more detailed usage, please refer to the official documentation.

Happy analyzing! ๐Ÿงฌโœจ