Merge pull request #1 from ArpiarSaundersLab/dev

Extended functionality added

Author: kennypavan

.gitignore

@@ -2,24 +2,16 @@ data/
 db/
 env/
 __pycache__/
-
-
+rdata/
 *.pyc
-
-# Setuptools distribution folder.
 /dist/
-
-# Python egg metadata, regenerated from source files by setuptools.
 /*.egg-info
 /*.egg
-
 env
 .eggs/
 build/
 build_doc.sh
 build_wheel.sh
-
 .DS_Store
-
 /**/__pycache__/
 /**/.pytest_cache/

examples/analysis/convert_to_seurat.R

@@ -0,0 +1,41 @@
+library(Seurat)
+library(SeuratDisk)
+library(anndata)
+
+# Define the file increments
+n_cells <- c("1000", "5000", "10000", "15000", "20000", "25000", "30000", "35000", "40000", "45000", "50000", "75000", "100000", "250000")
+
+#Iterate splatter data and convert to Seurat
+for (i in n_cells) {
+	print(i)
+	h5ad_file <- paste0("/home/kenny/Documents/OHSU/Projects/AnnSql/examples/data/splatter/data_", i, ".h5ad")
+	seurat_file <- paste0("/home/kenny/Documents/OHSU/Projects/AnnSql/examples/rdata/splatter/data_", i, ".rds")
+	Convert(h5ad_file, dest = "h5seurat", overwrite = TRUE)
+	seurat_obj <- LoadH5Seurat(sub(".h5ad$", ".h5seurat", h5ad_file))
+	saveRDS(seurat_obj, file = seurat_file)
+	file.remove(sub(".h5ad$", ".h5seurat", h5ad_file))
+	seurat_obj <- readRDS(seurat_file)
+	s4_seurat <- UpdateSeuratObject(object = seurat_obj)
+	saveRDS(s4_seurat, seurat_file)
+	rm(seurat_obj)
+	rm(s4_seurat)
+}
+
+
+# Iterate random data and convert to Seurat
+for (i in n_cells) {
+	print(i)
+	h5ad_file <- paste0("/home/kenny/Documents/OHSU/Projects/AnnSql/examples/data/random/data_", i, ".h5ad")
+	seurat_file <- paste0("/home/kenny/Documents/OHSU/Projects/AnnSql/examples/rdata/random/data_", i, ".rds")
+	Convert(h5ad_file, dest = "h5seurat", overwrite = TRUE)
+	seurat_obj <- LoadH5Seurat(sub(".h5ad$", ".h5seurat", h5ad_file), meta.data = FALSE, misc = FALSE)
+	adata <- read_h5ad(h5ad_file, backed = TRUE)
+	seurat_obj <- AddMetaData(seurat_obj, adata$obs)
+	saveRDS(seurat_obj, file = seurat_file)
+	file.remove(sub(".h5ad$", ".h5seurat", h5ad_file))
+	seurat_obj <- readRDS(seurat_file)
+	s4_seurat <- UpdateSeuratObject(object = seurat_obj)
+	saveRDS(s4_seurat, seurat_file)
+	rm(seurat_obj)
+	rm(s4_seurat)
+}

examples/analysis/generate_data.py examples/analysis/generate_data_random.pyexamples/analysis/generate_data.py renamed to examples/analysis/generate_data_random.py

@@ -29,16 +29,16 @@
 	adata.obs["cell_type"] = np.random.choice(["A","B","C"], adata.shape[0])
 
 	#save the object
-	adata.write("../data/random_data_"+str(i)+".h5ad")
+	adata.write("../data/random/data_"+str(i)+".h5ad")
 
 	#open the object
-	adata = sc.read("../data/random_data_"+str(i)+".h5ad", backed="r")
+	adata = sc.read("../data/random/data_"+str(i)+".h5ad", backed="r")
 
 	#make the database using backed mode
 	start_time = time.time()
-	MakeDb(adata, db_path="../db/", db_name="random_data_"+str(i), create_all_indexes=False, convenience_view=False)
+	MakeDb(adata, db_path="../db/random/", db_name="random_data_"+str(i), create_all_indexes=False, convenience_view=False)
 	print("--- %s seconds ---" % (time.time() - start_time))
 
 	#clear mem
 	gc.collect()
-	adata = None
+	adata = None

examples/analysis/generate_data_splatter.py

@@ -0,0 +1,63 @@
+#import libraries
+import scanpy as sc
+from AnnSQL import AnnSQL
+import time
+import pandas as pd
+from AnnSQL.MakeDb import MakeDb
+import os
+import numpy as np
+import gc
+import seaborn as sns
+import matplotlib.pyplot as plt
+from scsim import scsim
+
+
+#dataset sizes to generate
+dataset_sizes = [1000, 5000, 10000, 15000, 20000, 25000, 30000, 35000, 40000, 45000, 50000, 75000, 100000, 250000]
+dataset_sizes = [9999]
+
+#iterates and creates data_sizes defined above
+for i in dataset_sizes:
+
+	print(f"Running for {i}")
+
+	columns = 10000
+
+	simulator = scsim(ngenes=columns, ncells=i, ngroups=5, libloc=7.64, libscale=0.78,
+					mean_rate=7.68,mean_shape=0.34, expoutprob=0.00286,
+					expoutloc=6.15, expoutscale=0.49,
+					diffexpprob=0.025, diffexpdownprob=0., diffexploc=1.0, diffexpscale=1.0,
+					bcv_dispersion=0.448, bcv_dof=22.087, ndoublets=0,
+					nproggenes=400, progdownprob=0., progdeloc=1.0,
+					progdescale=1.0, progcellfrac=0.35, proggoups= list(range(1, int(10/3)+1)),
+					minprogusage=.1, maxprogusage=.7, seed=21)
+
+
+	start_time = time.time()
+	simulator.simulate()
+	end_time = time.time()
+	print('minutes elapsed for seed:',((end_time-start_time)/60))
+
+	#generate a cellxgene object using random data
+	adata = sc.AnnData(X=simulator.counts, 
+						obs=simulator.cellparams, 
+						)
+	adata.var.index = [f"gene_{i}" for i in range(adata.shape[1])]
+	adata.obs["cell_type"] = adata.obs["group"]
+	adata.obs = adata.obs.drop(columns=["group","has_program","program_usage","libsize"])
+
+	#save the object
+	adata.write("../data//splatter/data_"+str(i)+".h5ad")
+
+	#open the object
+	adata = sc.read("../data/splatter/data_"+str(i)+".h5ad", backed="r")
+
+	#make the database using backed mode
+	os.system(f"rm ../db/splatter/data_{i}.asql")
+	start_time = time.time()
+	MakeDb(adata, db_path="../db/splatter/", db_name="splatter_data_"+str(i), create_all_indexes=False, convenience_view=False)
+	print("--- %s seconds ---" % (time.time() - start_time))
+
+	#clear mem
+	gc.collect()
+	adata = None