Merge pull request #76 from aertslab/genome_object

Genome class to handle fasta files and chromsizes throughout package
aertslab · Dec 10, 2024 · 955bb1d · 955bb1d
2 parents d730793 + 01a46e7
commit 955bb1d
Show file tree

Hide file tree

Showing 23 changed files with 1,438 additions and 1,101 deletions.
diff --git a/docs/api/datasets.md b/docs/api/datasets.md
@@ -12,4 +12,6 @@ Downloading of use case datasets which ar explored in the example analyses.
 
     get_dataset
     get_motif_db
+    Genome
+    register_genome
 ```
diff --git a/docs/references.bib b/docs/references.bib
@@ -1,12 +1,3 @@
-@article{hu2023single,
-	title={Single-cell multi-scale footprinting reveals the modular organization of DNA regulatory elements},
-	author={Hu, Yan and Ma, Sai and Kartha, Vinay K and Duarte, Fabiana M and Horlbeck, Max and Zhang, Ruochi and Shrestha, Rojesh and Labade, Ajay and Kletzien, Heidi and Meliki, Alia and others},
-	journal={bioRxiv},
-	pages={2023--03},
-	year={2023},
-	publisher={Cold Spring Harbor Laboratory}
-}
-
 @misc{shrikumar2021tfmodisco,
   author       = {Av Shrikumar and Katherine Tian and annashcherbina and Žiga Avsec and Amr and Charles McAnany and pgreenside and Surag Nair and mhfzsharmin and Stefan Holderbach and Rosa Ma},
   title        = {{kundajelab/tfmodisco: Nicer API for density-adaptive hit scoring (v0.5.14.1)}},
@@ -26,3 +17,16 @@ @article{Virshup_2023
 	title = {The scverse project provides a computational ecosystem for single-cell omics data analysis},
 	journal = {Nature Biotechnology}
 }
+
+@article{Zhang_2024,
+    doi = {10.1038/s41592-023-02139-9},
+    url = {https://doi.org/10.1038/s41592-023-02139-9},
+    year = 2024,
+    month = {mar},
+    publisher = {Springer Nature},
+    author = {Kai Zhang and Nicholas R. Zemke and Evan J. Armand and others},
+    title = {A fast, scalable and versatile tool for analysis of single-cell omics data},
+    journal = {Nature Methods},
+    volume = {21},
+    pages = {217--227}
+}
diff --git a/docs/tutorials/enhancer_code_analysis.ipynb b/docs/tutorials/enhancer_code_analysis.ipynb
@@ -33,7 +33,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "metadata": {},
    "outputs": [
     {
@@ -54,9 +54,12 @@
     "adata = anndata.read_h5ad(\"mouse_biccn_data_filtered.h5ad\")\n",
     "\n",
     "genome_file = \"/home/VIB.LOCAL/niklas.kempynck/nkemp/software/dev_DeepPeak/DeepPeak/data/raw_mm/genome.fa\"\n",
+    "\n",
+    "genome = crested.Genome(genome_file)\n",
+    "\n",
     "datamodule = crested.tl.data.AnnDataModule(\n",
     "    adata,\n",
-    "    genome_file=genome_file,\n",
+    "    genome,\n",
     ")"
    ]
   },
@@ -315,7 +318,7 @@
    ],
    "source": [
     "%matplotlib inline\n",
-    "top_k=1000\n",
+    "top_k = 1000\n",
     "crested.pl.patterns.modisco_results(\n",
     "    classes=[\"Astro\", \"L5ET\", \"Vip\", \"Oligo\"],\n",
     "    contribution=\"positive\",\n",
@@ -691,7 +694,10 @@
     "    verbose=True,  # Useful for doing sanity checks on matching patterns\n",
     ")\n",
     "pattern_matrix = crested.tl.modisco.create_pattern_matrix(\n",
-    "    classes=list(adata.obs_names), all_patterns=all_patterns, normalize=False, pattern_parameter='seqlet_count_log'\n",
+    "    classes=list(adata.obs_names),\n",
+    "    all_patterns=all_patterns,\n",
+    "    normalize=False,\n",
+    "    pattern_parameter=\"seqlet_count_log\",\n",
     ")\n",
     "pattern_matrix.shape"
    ]
@@ -722,13 +728,21 @@
    ],
    "source": [
     "import matplotlib\n",
+    "\n",
     "%matplotlib inline\n",
-    "matplotlib.rcParams['pdf.fonttype'] = 42\n",
-    "matplotlib.rcParams['ps.fonttype'] = 42\n",
+    "matplotlib.rcParams[\"pdf.fonttype\"] = 42\n",
+    "matplotlib.rcParams[\"ps.fonttype\"] = 42\n",
     "\n",
     "pat_seqs = crested.tl.modisco.generate_nucleotide_sequences(all_patterns)\n",
     "crested.pl.patterns.clustermap(\n",
-    "    pattern_matrix, list(adata.obs_names), figsize=(16, 4.2), pat_seqs=pat_seqs, grid=True, fig_path='paperfigs/motif_clustering.pdf', dendrogram_ratio=(0.03,0.15), importance_threshold=4.5\n",
+    "    pattern_matrix,\n",
+    "    list(adata.obs_names),\n",
+    "    figsize=(16, 4.2),\n",
+    "    pat_seqs=pat_seqs,\n",
+    "    grid=True,\n",
+    "    fig_path=\"paperfigs/motif_clustering.pdf\",\n",
+    "    dendrogram_ratio=(0.03, 0.15),\n",
+    "    importance_threshold=4.5,\n",
     ")"
    ]
   },
@@ -791,12 +805,12 @@
     "crested.pl.patterns.clustermap(\n",
     "    pattern_matrix,\n",
     "    classes=list(adata.obs_names),\n",
-    "    subset=['L2_3IT', 'L5ET', 'L5IT', 'L5_6NP', 'L6CT', 'L6IT','L6b'],\n",
+    "    subset=[\"L2_3IT\", \"L5ET\", \"L5IT\", \"L5_6NP\", \"L6CT\", \"L6IT\", \"L6b\"],\n",
     "    figsize=(10, 2),\n",
     "    pat_seqs=pat_seqs,\n",
     "    grid=True,\n",
     "    dy=0.0025,\n",
-    "    importance_threshold=4.5\n",
+    "    importance_threshold=4.5,\n",
     ")"
    ]
   },
@@ -1022,6 +1036,7 @@
    ],
    "source": [
     "import crested\n",
+    "\n",
     "file_path = \"/home/VIB.LOCAL/niklas.kempynck/nkemp/mouse/biccn/Mouse_rna.h5ad\"  # Locate h5 file containing scRNAseq data\n",
     "cell_type_column = \"subclass_Bakken_2022\"\n",
     "mean_expression_df = crested.tl.modisco.calculate_mean_expression_per_cell_type(\n",
@@ -1046,7 +1061,9 @@
     }
    ],
    "source": [
-    "crested.pl.patterns.tf_expression_per_cell_type(mean_expression_df, ['Nfia', 'Spi1', 'Mef2c'])"
+    "crested.pl.patterns.tf_expression_per_cell_type(\n",
+    "    mean_expression_df, [\"Nfia\", \"Spi1\", \"Mef2c\"]\n",
+    ")"
    ]
   },
   {
@@ -1486,7 +1503,7 @@
     "    normalize_gex=True,\n",
     "    min_tf_gex=0.95,\n",
     "    importance_threshold=5,\n",
-    "    pattern_parameter='seqlet_count_log',\n",
+    "    pattern_parameter=\"seqlet_count_log\",\n",
     "    filter_correlation=True,\n",
     "    verbose=True,\n",
     "    zscore_threshold=1,\n",
@@ -1528,13 +1545,13 @@
    "source": [
     "crested.pl.patterns.clustermap_tf_motif(\n",
     "    tf_ct_matrix,\n",
-    "    heatmap_dim='contrib',\n",
-    "    dot_dim='gex',\n",
+    "    heatmap_dim=\"contrib\",\n",
+    "    dot_dim=\"gex\",\n",
     "    class_labels=classes,\n",
     "    pattern_labels=tf_pattern_annots,\n",
-    "    fig_size=(35,6),\n",
+    "    fig_size=(35, 6),\n",
     "    cluster_rows=True,\n",
-    "    cluster_columns=False\n",
+    "    cluster_columns=False,\n",
     ")"
    ]
   },