import networkx as nx
import pandas as pd
from IPython.display import Image, display
import scphylo as scp
from scphylo.pl._helper import (
_add_barplot,
_add_chromplot,
_clonal_cell_mutation_list,
_get_tree,
_newick_info2_mutation_list,
)
[docs]def clonal_tree(
tree,
muts_as_number=False,
cells_as_number=False,
show_id=False,
cell_info=None,
output_file=None,
color_attr=None,
dpi=150,
):
"""Draw the tree in clonal format.
This functions plots the tree in which edges are mutations and nodes are cells.
Parameters
----------
tree : :class:`networkx.DiGraph`
The input tree.
muts_as_number : :obj:`bool`, optional
Change the mutation list to a number at edges, by default False
cells_as_number : :obj:`bool`, optional
Change the cell list to a number at edges, by default False
show_id : :obj:`bool`, optional
Whether to show IDs of nodes and edges or not, by default True
cell_info : :class:`pandas.DataFrame`, optional
Information of cells for coloring the nodes by a pie chart, by default None
output_file : :obj:`str`, optional
Path to a file for saving the tree in, by default None
color_attr : :obj:`str`, optional
Attributes in the `cell_info` dataframe for coloring the nodes, by default None
dpi : :obj:`int`, optional
Resolution of rendered figures – this influences the size of
figures in notebooks, by default 150
Returns
-------
:obj:`None`
"""
_, graphviz_is_not_imporeted = scp.ul.import_graphviz()
if graphviz_is_not_imporeted:
scp.logg.error("Unable to import a package!")
tc = tree.copy()
root = scp.ul.root_id(tree)
tc.nodes[root]["label"] = "root"
tc.nodes[root]["fontname"] = "Helvetica"
tc.nodes[root]["style"] = "rounded"
tc.nodes[root]["shape"] = "box"
tc.nodes[root]["margin"] = 0.05
tc.nodes[root]["pad"] = 0
tc.nodes[root]["width"] = 0
tc.nodes[root]["height"] = 0
if muts_as_number:
for u, v, label in tc.edges.data("label"):
if label == "":
ll = []
else:
ll = label.split(tc.graph["splitter_mut"])
tc.add_edge(u, v, label=f" {len(ll)} ")
if cells_as_number:
for n in tc.nodes:
if n != root:
ll = tc.nodes[n]["label"].split(tc.graph["splitter_cell"])
if "––" in tc.nodes[n]["label"]:
tc.nodes[n]["label"] = "0"
else:
tc.nodes[n]["label"] = f"{len(ll)}"
if cell_info is not None:
tc.nodes[root]["label"] = tree.graph["splitter_cell"].join(
tc.graph["normal_cells"]
)
mapping = cell_info[color_attr].to_dict()
for node in tc:
num = 0
paths = nx.shortest_path(tc, source=root, target=node)
for i in range(len(paths) - 1):
x = paths[i]
y = paths[i + 1]
num += len(tc[x][y]["label"].split(tc.graph["splitter_mut"]))
try:
freq = [
mapping[x]
for x in tc.nodes[node]["label"].split(tc.graph["splitter_cell"])
]
except KeyError:
freq = ["#FFFFFF"]
freq = pd.DataFrame(freq)[0].value_counts(normalize=True)
fillcolor = ""
for index, value in freq.items():
fillcolor += f"{index};{value}:"
tc.nodes[node]["fontsize"] = 14
tc.nodes[node]["shape"] = "circle"
tc.nodes[node]["fontname"] = "Helvetica"
tc.nodes[node]["style"] = "wedged"
tc.nodes[node]["margin"] = 0.05
tc.nodes[node]["pad"] = 0
if "––" in tc.nodes[node]["label"]:
tc.nodes[node]["width"] = 0
tc.nodes[node]["height"] = 0
else:
tc.nodes[node]["width"] = 0.8
tc.nodes[node]["height"] = 0.8
tc.nodes[node]["fillcolor"] = fillcolor # "red;0.3:green;0.6:orange"
tc.nodes[node]["label"] = len(
tc.nodes[node]["label"].split(tc.graph["splitter_cell"])
)
tc.nodes[node]["fontcolor"] = "white"
tc.nodes[node]["color"] = "gray"
if show_id:
for u, v, label in tc.edges.data("label"):
tc.add_edge(u, v, label=label + f"\n[{v}]")
tc.nodes[v]["label"] = tc.nodes[v]["label"] + f"\n[{v}]"
tc.graph["graph"] = {"fontname": "Helvetica"}
tc.graph["node"] = {"fontname": "Helvetica", "fontsize": 14}
tc.graph["edge"] = {"fontname": "Helvetica", "fontsize": 14}
tree.graph["type"] = "clonal"
cell_list, mutation_list = _clonal_cell_mutation_list(tree)
tree.graph["cell_list"] = cell_list
tree.graph["mutation_list"] = mutation_list
if output_file is not None:
scp.io.to_png(tc, output_file, dpi)
return display(
Image(nx.drawing.nx_pydot.to_pydot(tc).create_png(), embed=True, retina=True)
)
[docs]def dendro_tree(
tree,
width=1200,
height=500,
dpi=300,
cell_info=None,
label_color="group_color",
line_size=0.4,
tiplab_size=2.5,
inner_node_type="nmuts",
inner_node_size=2,
distance_labels_to_bottom=4,
annotation=None,
output_file=None,
):
"""Draw the tree in dendro fromat.
Parameters
----------
tree : :class:`networkx.DiGraph`
The input tree.
width : :obj:`int`, optional
Width of the figure, by default 8
height : :obj:`int`, optional
Height of the figure, by default 3
dpi : :obj:`int`, optional
The resolution, by default 300
cell_info : :class:`pandas.DataFrame`, optional
Information about cells such as color,
expression values of genes and etc, by default None
label_color : :obj:`str`, optional
The column name in which colors of cells are stored
in the dataframe provided as `cell_info`, by default "group_color"
line_size : :obj:`float`, optional
Line size of the tree, by default 0.4
tiplab_size : :obj:`float`, optional
Cell name size in the tree, by default 2.5
inner_node_type : :obj:`str`, optional
The format of the inner nodes (i.e. mutations), by default "nmuts"
Values are:
- `nmuts`: only number of mutations
- `nodeid`: only node id
- `both`: both number of mutations and node id
inner_node_size : :obj:`int`, optional
Size of the inner nodes (i.e. mutations), by default 2
distance_labels_to_bottom : :obj:`int`, optional
Distance cell names to the bottom of the figure, by default 4
annotation : :obj:`list`, optional
List of gene names provided in the column dataframe of ``cell_info``
in to be annotated in the bottom of the tree, by default []
output_file : :obj:`str`, optional
Path to a file for saving the tree in, by default None
Returns
-------
:obj:`None`
Note
----
The cell names in the tree must be identical to the index of `cell_info`
dataframe if it was provided.
"""
if annotation is None:
annotation = []
if inner_node_type.lower() not in ["nmuts", "nodeid", "both"]:
scp.logg.error("Wrong `inner_node_type` choice!")
ggtree, ggtree_is_not_imported = scp.ul.import_rpy2(
"ggtree",
"devtools::install_github(c('YuLab-SMU/ggtree','xiangpin/ggtreeExtra'"
+ ",'YuLab-SMU/aplot'))\ninstall.packages('cowplot')\n",
)
if ggtree_is_not_imported:
scp.logg.error("Unable to import a package!")
import rpy2.robjects as ro
from rpy2.robjects import pandas2ri
from rpy2.robjects.lib import grdevices
from rpy2.robjects.packages import importr
newick, info2, mutation_list = _newick_info2_mutation_list(tree)
tree.graph["mutation_list"] = mutation_list.set_index("index")
tree.graph["newick"] = newick
tree.graph["type"] = "dendro"
importr("ggplot2")
importr("cowplot")
importr("ggtree")
importr("ape")
importr("aplot")
with ro.conversion.localconverter(ro.default_converter + pandas2ri.converter):
if cell_info is not None:
if "group_color" not in cell_info.columns:
cell_info["group_color"] = tree.graph["data"].shape[0] * ["#000000"]
cell_info_r = ro.conversion.py2rpy(cell_info.reset_index())
else:
cell_info_r = ro.conversion.py2rpy(
pd.DataFrame.from_dict(
{
"x": tree.graph["data"].index,
"group_color": tree.graph["data"].shape[0] * ["#000000"],
}
)
)
ro.globalenv["info1"] = cell_info_r
info2_r = ro.conversion.py2rpy(info2)
ro.globalenv["info2"] = info2_r
cmd = _get_tree(
newick,
line_size,
label_color,
tiplab_size,
inner_node_type,
inner_node_size,
distance_labels_to_bottom,
)
for i, ann in enumerate(annotation):
if i == 0:
cmd += "info1$id_index <- p$data[match(info1[,1], p$data$label),]$y"
if ann[0] == "bar":
cmd += _add_barplot(ann[1], ann[2], ann[3])
elif ann == "chrom":
cmd += _add_chromplot()
elif ann == "imputation":
for imp in [
("f_0_1", "f_0_1_color"),
("f_1_0", "f_1_0_color"),
("f_3_0", "f_3_0_color"),
("f_3_1", "f_3_1_color"),
]:
cmd += _add_barplot(imp[0], imp[1])
with grdevices.render_to_bytesio(
grdevices.png, width=width, height=height, res=dpi
) as image:
p = ro.r(cmd)
ro.r.show(p)
if output_file is not None:
ro.r.ggsave(
plot=p,
filename=output_file,
width=width / dpi,
height=height / dpi,
units="in",
dpi=dpi,
limitsize=False,
)
return display(Image(image.getvalue(), embed=True, retina=True))
def newick_tree(
adata,
tree_attr="tree",
leaf_attr=None,
width=2000,
height=800,
dpi=300,
tippoint_size=0.8,
tiplab_size=0.9,
show_tiplab=False,
):
"""[summary].
Parameters
----------
adata : [type]
[description]
leaf_attr : [type]
[description]
tree_attr : [type]
[description]
width : int, optional
[description], by default 2000
height : int, optional
[description], by default 800
dpi : int, optional
[description], by default 300
tippoint_size : float, optional
[description], by default 0.8
tiplab_size : float, optional
[description], by default 0.9
show_tiplab : bool, optional
[description], by default False
Returns
-------
[type]
[description]
"""
ggtree, ggtree_is_not_imported = scp.ul.import_rpy2(
"ggtree",
"devtools::install_github(c('YuLab-SMU/ggtree','xiangpin/ggtreeExtra'"
+ ",'YuLab-SMU/aplot'))\ninstall.packages('cowplot')\n",
)
if ggtree_is_not_imported:
scp.logg.error("Unable to import a package!")
import rpy2.robjects as ro
from rpy2.robjects import pandas2ri
from rpy2.robjects.lib import grdevices
from rpy2.robjects.packages import importr
importr("ggplot2")
if leaf_attr is not None:
df = adata.obs[[leaf_attr]]
else:
df = adata.obs.copy()
df["color"] = "#000000"
df = df[["color"]]
leaf_attr = "color"
with ro.conversion.localconverter(ro.default_converter + pandas2ri.converter):
info_color_r = ro.conversion.py2rpy(df.reset_index())
ro.globalenv["info"] = info_color_r
cmd = f"""
tree <- read.tree(text='{str(adata.uns[tree_attr]).strip()}')
p <- ggtree(tree, layout='dendrogram', size=0.2, color='gray')
# p <- p %>% rotate(191) %>% rotate(218)
# p <- p + geom_text(aes(label=node), size=1)
p <- p %<+% info + geom_tippoint(aes(color={leaf_attr}), size={tippoint_size})
"""
if show_tiplab:
cmd += f"""
p <- p + geom_tiplab(aes(color={leaf_attr}), angle=90,
size={tiplab_size}, hjust=1.1)
"""
cmd += f"""
p <- p + scale_color_identity(guide='none')
# ggsave('/data/frashidi/{leaf_attr}.pdf', p, width=7.2, height=2,
# units='in', limitsize=FALSE, dpi=300)
p
"""
with grdevices.render_to_bytesio(
grdevices.png, width=width, height=height, res=dpi
) as image:
p = ro.r(cmd)
ro.r.show(p)
return display(Image(image.getvalue(), embed=True, retina=True))
def networkx_tree(tree, n_attr="label", e_attr="label"):
"""[summary].
Parameters
----------
tree : [type]
[description]
n_attr : str, optional
[description], by default "label"
e_attr : str, optional
[description], by default "label"
"""
graph2 = tree.copy()
if n_attr == "id":
for n in graph2.nodes():
if graph2.in_degree(n) == 1 and graph2.out_degree(n) >= 1:
graph2.nodes[n]["shape"] = "circle"
graph2.nodes[n]["margin"] = 0
graph2.nodes[n]["label"] = n
else:
for n in graph2.nodes():
if n_attr in graph2.nodes[n]:
graph2.nodes[n]["label"] = graph2.nodes[n][n_attr]
for e in graph2.edges():
if e_attr in graph2.edges[e]:
graph2.edges[e]["label"] = graph2.edges[e][e_attr]
display(Image(nx.drawing.nx_pydot.to_pydot(graph2).create_png(), retina=True))