Skip to content

Commit

Permalink
Placed each plot into its own set of if statements
Browse files Browse the repository at this point in the history
  • Loading branch information
@JeffreyMaurer
JeffreyMaurer committed Oct 8, 2024
1 parent ff0e222 commit 5737478
Showing 1 changed file with 98 additions and 73 deletions.
171 changes: 98 additions & 73 deletions scripts/05_Filter_Merged_Marker.py
View file
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
#!/usr/bin/env python3
# coding: utf-8

# JM 08/17/24 # This is going to have the merged
# JM 08/17/24
# In this notebook, we're going plot the mouse retinal data and determine proper marker genes.

import datetime
Expand Down Expand Up @@ -34,6 +34,8 @@
plot_minor_markers = True
plot_minor_cells = True
plot_split_majorclass = True
plot_only_target_cells = False
plot_only_curated = False

raw = True
xenium_filtered = True
Expand All @@ -48,15 +50,8 @@
if xenium_filtered:
data_string = data_string + "_xeniumFiltered"

plot_occassion = ""
plot_occassion = "" # Options: "august_grant": # 05.1, "curated_xenium_filtered": # 05.2
data_string = data_string + f"_{plot_occassion}"
markers = ""
if plot_occassion == "august_grant": # 05.1
pass
elif plot_occassion == "curated_xenium_filtered": # 05.2
pass
else:
pass

# Clean subtypes
adata.obs = adata.obs.loc[:, ["majorclass", "author_cell_type"]]
Expand Down Expand Up @@ -130,7 +125,7 @@

## Expression Filtering
merged_filtered_markers["Detectable_Expression"] = merged_filtered_markers["Major_Detectable_Expression"] | merged_filtered_markers["Minor_Detectable_Expression"]
merged_filtered_markers["not_crowding"] = np.sum(merged_filtered_markers.loc[:, 'AC':'Rod'] > 100, axis = 1) == 0
merged_filtered_markers["not_crowding"] = np.sum(merged_filtered_markers.loc[:, 'AC':'Rod'] > 100, axis = 1) == 0 # If subtype is high, that should be ok

# Filter
merged_filtered_markers["xenium_filter"] = (merged_filtered_markers["long_enough"] &
Expand All @@ -146,6 +141,10 @@
# merged_filtered_markers = pd.read_csv('05_Filter_Merged_Markers/5_curated_markers_annotatedKeep_lengthExpressionMissingFiltered.txt', sep ='\t')
# merged_filtered_markers = merged_filtered_markers.sort_values(['Queried_Major_Name', 'Queried_Name']) # For now

if plot_only_curated:
merged_filtered_markers.loc[merged_filtered_markers["Curated"] == "Curated"]
data_string = data_string + '_CuratedOnly'

def clean_markers(var_names, dirty_markers, verbose=True):
final_markers = []
for m in dirty_markers:
Expand All @@ -161,11 +160,10 @@ def clean_markers(var_names, dirty_markers, verbose=True):
return(final_markers)

# 05.2
if plot_major:
if plot_major_markers and plot_major_cells:
assume_correct = merged_filtered_markers["Queried_Major_Name"] == merged_filtered_markers["Queried_Name"]
hedge_unharmonized = merged_filtered_markers["Queried_Major_Name"] == merged_filtered_markers["Name"]
curated_majorclass_markers = merged_filtered_markers.loc[assume_correct | hedge_unharmonized]
curated_majorclass_markers.to_csv('05_Filter_Merged_Markers/5_curatedMarkers_majorclass_xeniumFiltered.txt', sep = '\t', index = False)

final_markers = clean_markers(adata.var_names, curated_majorclass_markers["Marker"].tolist())

Expand All @@ -177,74 +175,101 @@ def clean_markers(var_names, dirty_markers, verbose=True):
vmin = 0,
show = False,
save = f"/project/hipaa_ycheng11lab/atlas/CAMR2024/05_Filter_Merged_Markers/figures/5_dotplot_mouseRetina_majorclass_curatedMarkers_{data_string}.pdf")


for majorclass in adata.obs['majorclass'].cat.categories:
# End plot_major_markers and plot_major_cells

if plot_minor_markers and plot_minor_cells:
all_markers = []
all_target_subtypes = []
for majorclass in adata.obs['majorclass'].cat.categories:

print(f'{datetime.datetime.now()} Major Class: {majorclass}')

majorclass_original = majorclass
majorclass = majorclass.upper()

is_subtype_marker = np.logical_and(merged_filtered_markers["Major_Name"] == majorclass, merged_filtered_markers["Name"] != majorclass)
markers = merged_filtered_markers.loc[is_subtype_marker, "Marker"].tolist()
target_subtypes = merged_filtered_markers.loc[is_subtype_marker, "Queried_Name"].drop_duplicates(keep='first').tolist()

if plot_major_markers:
is_major_marker = np.logical_and(merged_filtered_markers["Major_Name"] == majorclass, merged_filtered_markers["Name"] == majorclass)
cell_markers = merged_filtered_markers.loc[is_major_marker, "Marker"].tolist()
markers = cell_markers + markers

final_markers = clean_markers(adata.var_names, markers)
all_markers += final_markers


if not plot_split_majorclass:
continue

if final_markers == []: # Necessary to avoid "ValueError: left cannot be >= right"
print(f'No minor markers available for {majorclass}!')
continue

if not plot_only_target_cells:
all_subtypes = adata.obs.loc[adata.obs['majorclass'] == majorclass_original, "author_cell_type"].astype(str).sort_values().drop_duplicates(keep='first')
target_subtypes += all_subtypes[!all_subtypes.isin(target_subtypes)].tolist()
all_target_subtypes += target_subtypes

sc.pl.dotplot(adata[adata.obs['majorclass'] == majorclass_original, final_markers],
var_names = final_markers,
groupby = 'author_cell_type',
categories_order = target_subtypes,
vmax = max_col,
vmin = 0,
show = False,
save = f"mouseRetina_minorclass-{majorclass}_{data_string}.pdf")
# End majorclass loop

print(f'{datetime.datetime.now()} Major Class: {majorclass}')
final_markers = clean_markers(adata.var_names, all_markers)
sc.pl.dotplot(adata[:, final_markers],
var_names = final_markers,
groupby = 'author_cell_type',
categories_order = all_target_subtypes,
vmax = max_col,
vmin = 0,
show = False,
save = f"mouseRetina_minorclass_{data_string}.pdf")
# End plot_minor_markers and plot_minor_cells

if plot_occassion == "august_grant": # 05.1
# Make majorclass in author_cell_type uppercase to match Name and Major_Name
majorclass_idx = adata.obs["author_cell_type"].str.upper().isin(adata.obs["majorclass"].astype(str).str.upper())
adata.obs.loc[majorclass_idx, "author_cell_type"] = adata.obs.loc[majorclass_idx, "author_cell_type"].str.upper()

majorclass_original = majorclass
majorclass = majorclass.upper()
merged_filtered_markers_Curated = merged_filtered_markers.loc[merged_filtered_markers["Curated"] == "Curated"]
merged_filtered_markers_AC = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "AC"]
merged_filtered_markers_BC = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "BC"]
merged_filtered_markers_RGC = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "RGC"]
merged_filtered_markers_MG = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "MG"]
merged_filtered_markers_RPE = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "RPE"]
merged_filtered_markers_Curated = pd.concat([merged_filtered_markers_AC,
merged_filtered_markers_BC,
merged_filtered_markers_RGC,
merged_filtered_markers_MG,
merged_filtered_markers_RPE],
ignore_index = True)

is_major_marker = np.logical_and(merged_filtered_markers["Major_Name"] == majorclass, merged_filtered_markers["Name"] == majorclass)
cell_markers = merged_filtered_markers.loc[is_major_marker, "Marker"].tolist()
ordered_celltypes = merged_filtered_markers_Curated["Queried_Name"].drop_duplicates().astype('category').cat.remove_categories('RGC').dropna().tolist() + ["ROD","CONE","HC","MICROGLIA","ENDOTHELIAL"]

is_subtype_marker = np.logical_and(merged_filtered_markers["Major_Name"] == majorclass, merged_filtered_markers["Name"] != majorclass)
subtype_markers = merged_filtered_markers.loc[is_subtype_marker, "Marker"].tolist()
august_grant_markers = ["Ptn","Cntn6","Nxph1","Cpne4","Cbln4","Etv1","Epha3","Trhde", # AC
"Neto1","Erbb4","Grik1","Nnat","Cabp5","Sox6","Cck","Cpne9","Prkca","Hcn1", # BC
"Rbpms","Penk","Spp1","Coch","Opn4","Il1rapl2","Tbx20","Tac1","Cartpt", # RGC
"Aqp4","Glul","Rlbp1","Slc1a3","Vim", # MG
"Rpe65", # RPE
"Rho", "Arr3","Lhx1","Onecut1", "Cd74", "Cldn5"] # Manual

markers = cell_markers + subtype_markers

final_markers = clean_markers(adata.var_names, markers)

if final_markers == []: # Necessary to avoid "ValueError: left cannot be >= right"
print(f'No markers available for {majorclass}!')
continue
# Temporary change for this plot
adata.obs.loc[adata.obs["author_cell_type"].astype(str).str.contains("Microglia"), "author_cell_type"] = "MICROGLIA" # No subtypes of Microglia here

sc.pl.dotplot(adata[adata.obs['majorclass'] == majorclass_original, final_markers],
var_names = final_markers,
sc.pl.dotplot(adata[adata.obs['author_cell_type'].isin(ordered_celltypes), :],
var_names = august_grant_markers,
groupby = 'author_cell_type',
categories_order = adata.obs.loc[adata.obs['majorclass'] == majorclass_original, "author_cell_type"].astype(str).sort_values().drop_duplicates(keep='first').tolist(), # Only celltypes that have a marker should be present
categories_order = ordered_celltypes,
vmax = max_col,
vmin = 0,
show = False,
save = f"mouseRetina_minorclass-{majorclass}_filteredMergedMarkers_{data_string}.pdf")
# End majorclass

# Make majorclass in author_cell_type uppercase to match Name and Major_Name
majorclass_idx = adata.obs["author_cell_type"].str.upper().isin(adata.obs["majorclass"].astype(str).str.upper())
adata.obs.loc[majorclass_idx, "author_cell_type"] = adata.obs.loc[majorclass_idx, "author_cell_type"].str.upper()

merged_filtered_markers_Curated = merged_filtered_markers.loc[merged_filtered_markers["Curated"] == "Curated"]
merged_filtered_markers_AC = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "AC"]
merged_filtered_markers_BC = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "BC"]
merged_filtered_markers_RGC = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "RGC"]
merged_filtered_markers_MG = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "MG"]
merged_filtered_markers_RPE = merged_filtered_markers_Curated.loc[merged_filtered_markers_Curated["Major_Name"] == "RPE"]
merged_filtered_markers_Curated = pd.concat([merged_filtered_markers_AC,
merged_filtered_markers_BC,
merged_filtered_markers_RGC,
merged_filtered_markers_MG,
merged_filtered_markers_RPE],
ignore_index = True)

ordered_celltypes = merged_filtered_markers_Curated["Queried_Name"].drop_duplicates().astype('category').cat.remove_categories('RGC').dropna().tolist() + ["ROD","CONE","HC","MICROGLIA","ENDOTHELIAL"]

august_grant_markers = ["Ptn","Cntn6","Nxph1","Cpne4","Cbln4","Etv1","Epha3","Trhde", # AC
"Neto1","Erbb4","Grik1","Nnat","Cabp5","Sox6","Cck","Cpne9","Prkca","Hcn1", # BC
"Rbpms","Penk","Spp1","Coch","Opn4","Il1rapl2","Tbx20","Tac1","Cartpt", # RGC
"Aqp4","Glul","Rlbp1","Slc1a3","Vim", # MG
"Rpe65", # RPE
"Rho", "Arr3","Lhx1","Onecut1", "Cd74", "Cldn5"] # Manual

# Temporary change for this plot
adata.obs.loc[adata.obs["author_cell_type"].astype(str).str.contains("Microglia"), "author_cell_type"] = "MICROGLIA" # No subtypes of Microglia here

sc.pl.dotplot(adata[adata.obs['author_cell_type'].isin(ordered_celltypes), :],
var_names = august_grant_markers,
groupby = 'author_cell_type',
categories_order = ordered_celltypes,
vmax = max_col,
vmin = 0,
show = False,
figsize = (12,10),
save = f"mouseRetina_filteredMergedMarkers_curated_{data_string}.pdf")
figsize = (12,10),
save = f"mouseRetina_curated_{data_string}.pdf")
# End August Grant

0 comments on commit 5737478

Please sign in to comment.