def allvlmc_vs_allfb__mrkr():
ab = df_expr_ab.reindex(mrkr.index, axis=1).dropna(axis=1)
bh = df_expr_bh.reindex(mrkr.index, axis=1).dropna(axis=1)
assert set(ab.columns) == set(bh.columns)
ab = df_expr_ab[df_meta_ab.subclass_label == "VLMC"]
bh = df_expr_bh[df_meta_bh.celltype.str.startswith("FB")]
ab = ab.reindex(
df_meta_ab.reindex(ab.index).cell_type_alias_label.sort_values().index)
bh = bh.reindex(df_meta_bh.reindex(bh.index).celltype.sort_values().index)
S = pd.DataFrame(index=bh.index,
columns=ab.index,
data=cosine_similarity(bh, ab))
njj = ab.index.groupby(
df_meta_ab.reindex(ab.index).cell_type_alias_label.to_list())
mii = bh.index.groupby(df_meta_bh.reindex(bh.index).celltype.to_list())
fig: plt.Figure
(fig, AX) = plt.subplots(len(mii), len(njj))
for (i, (m, ii)) in enumerate(mii.items()):
for (j, (n, jj)) in enumerate(njj.items()):
ax = AX[i, j]
ax.imshow(S.loc[ii, jj], aspect='auto', cmap='Greys')
ax.axis('off')
# No effect?
plt.xlabel("Mouse-WCH-2020")
plt.ylabel("Betsholtz-2018")
fig.savefig(out_dir / f"{whatsmyname()}.png")
def avg_cos_sim__mrkr():
ab = df_expr_ab.reindex(mrkr.index, axis=1).dropna(axis=1)
bh = df_expr_bh.reindex(mrkr.index, axis=1).dropna(axis=1)
assert set(ab.columns) == set(bh.columns)
# # DEBUG
# df_expr_bh = df_expr_bh.sample(n=9, random_state=43)
# df_expr_ab = df_expr_ab.sample(n=8, random_state=43)
# Cell label
anno_bh = df_meta_bh.reindex(bh.index).celltype
anno_ab = df_meta_ab.reindex(ab.index).cell_type_alias_label
bh = bh.groupby(by=list(anno_bh)).mean()
ab = ab.groupby(by=list(anno_ab)).mean()
S = pd.DataFrame(index=bh.index,
columns=ab.index,
data=cosine_similarity(bh, ab))
with Plox() as px:
px.a.imshow(S, cmap="Greys")
px.a.set_xticks(range(len(S.columns)))
px.a.set_xticklabels(S.columns, rotation=90)
px.a.set_xlabel("Mouse-WCH-2020")
px.a.set_yticks(range(len(S.index)))
px.a.set_yticklabels(S.index)
px.a.set_ylabel("Betsholtz-2018")
px.f.savefig(out_dir / f"{whatsmyname()}.png")
def avg_cos_sim__mrkr():
# This is from Betsholtz-2018
# https://www.nature.com/articles/nature25739/figures/1
# 1c
markers = {
'PC': "Pdgfrb Cspg4 Anpep Rgs5 Cd248 Abcc9 Vtn S1pr3",
'SMC': "Acta2 Tagln Myh11 Myl9 Mylk Sncg Cnn1 Pln",
'MG':
"Csf1r Cd68 Cd53 Cd48 Cd84 C1qa Fcgr1", # removed Emr1 (missing in AB)
'FB': "Pdgfra Lum Dcn Col3a1 Col5a1 Col8a2 Col12a1 Mmp2",
'OL': "Mobp Plp1 Mog Cldn11 Mag Gjc2 Mal Cnp",
'EC': "Pecam1 Kdr Flt1 Tie1 Tek Icam2 Podxl Ptprb",
'AC': "Aldh1l1 Fgfr3 Slc4a4 Slc6a11 Slc7a10 Mlc1 Slc1a3 Cldn10",
}
markers = sorted(set(' '.join(markers.values()).split(' ')))
ab = df_expr_ab.loc[markers].T
cb = df_expr_cb.loc[markers].T
assert set(ab.columns) == set(cb.columns)
# # DEBUG
# df_expr_cb = df_expr_cb.sample(n=9, random_state=43)
# df_expr_ab = df_expr_ab.sample(n=8, random_state=43)
# Cell label
anno_cb = df_meta_cb.reindex(cb.index).celltype
anno_ab = df_meta_ab.reindex(ab.index).cell_type_alias_label
cb = cb.groupby(by=list(anno_cb)).mean()
ab = ab.groupby(by=list(anno_ab)).mean()
S = pd.DataFrame(index=cb.index,
columns=ab.index,
data=cosine_similarity(cb, ab))
with Plox() as px:
px.a.imshow(S, cmap="Greys")
px.a.set_xticks(range(len(S.columns)))
px.a.set_xticklabels(S.columns, rotation=90)
px.a.set_xlabel("Mouse-WCH-2020")
px.a.set_yticks(range(len(S.index)))
px.a.set_yticklabels(S.index)
px.a.set_ylabel(dataset_name)
px.f.savefig(out_dir / f"{whatsmyname()}.png")
def cluster(df_expr):
df = pd.DataFrame(index=df_expr.index,
columns=["x", "y"],
data=TSNE(random_state=43).fit_transform(df_expr))
df = df.assign(celltype=list(df_meta.reindex(df.index).celltype))
return df
from sklearn.manifold import TSNE
out_dir = mkdir(Path(__file__).with_suffix(''))
# df_expr = pd.DataFrame(np.random.RandomState(0).random(size=(30, 5)))
from z_sources import df_expr, df_meta, df_mrkr
# # DEBUG -- subset samples
# df_expr = df_expr.sample(n=100, random_state=43)
# # Subset to marker genes
# df_expr = df_expr[df_mrkr.index]
# Order samples
df_meta = df_meta.reindex(df_expr.index)
styles = pd.DataFrame(index=['marker', 'c'],
data={
'PC': ('s', 'red'),
'vSMC': ('s', 'green'),
'aaSMC': ('o', 'green'),
'aSMC': ('^', 'green'),
'MG': ('s', 'gray'),
'FB1': ('^', 'purple'),
'FB2': ('s', 'violet'),
'OL': ('s', 'brown'),
'EC1': ('s', 'cyan'),
'EC2': ('o', 'cyan'),
'EC3': ('^', 'cyan'),
'vEC': ('s', 'blue'),
def histograms():
genes = [
"Trpm3",
"mt-Co1",
"mt-Co3",
"Nnat",
"Ptgds",
"Adam12",
"Alcam", # Up early
"Itih5",
"Malat1",
"Zbtb20",
"Spp1",
"Col15a1",
"Ece1",
"Cemip", # Up late
]
ab = df_expr_ab.reindex(genes, axis=1).dropna(axis=1)
bh = df_expr_bh.reindex(genes, axis=1).dropna(axis=1)
assert list(ab.columns) == list(bh.columns)
# Norm1 normalization
# ab = ab.div(ab.sum(axis=1), axis=0)
# bh = bh.div(bh.sum(axis=1), axis=0)
# log1p trafo
ab = ab.transform(lambda x: np.log(x + 1))
bh = bh.transform(lambda x: np.log(x + 1))
ab = ab[df_meta_ab.subclass_label == "VLMC"]
bh = bh[df_meta_bh.celltype.str.startswith("FB")]
assert list(ab.columns) == list(bh.columns)
genes = sorted(ab.columns)
# g = first(genes)
colors = {
'FB1': 'purple',
'FB2': 'violet',
'374_VLMC': "C0",
'375_VLMC': "C1",
'376_VLMC': "C2",
}
for g in genes:
expr: pd.Series
with Plox() as px:
grps = [
ab[g].groupby(
df_meta_ab.reindex(ab.index).cell_type_alias_label),
bh[g].groupby(df_meta_bh.reindex(bh.index).celltype),
]
for grp in grps:
for (label, expr) in grp:
if any(expr):
f = gaussian_kde(expr)
xx = np.linspace(0, max(expr) * 1.5, 100)
px.a.plot(xx,
f(xx),
label=f"{label} ({len(expr)})",
color=colors[label])
px.a.set_xlabel("log1p(count)")
px.a.legend()
px.a.set_yticks([])
px.f.savefig(out_dir / f"hist_{g}.png")