Quickstart: Build and Inspect an AnnNet¶

This notebook introduces the core object, annnet.AnnNet.

You will learn how to:

• create a graph
• add vertices, binary edges, and hyperedges
• inspect the incidence matrix and attribute tables
• read the high-level graph summary fields exposed by the API

In [1]:

from pathlib import Path
import sys

NOTEBOOK_DIR = Path.cwd()
PACKAGE_ROOT = (NOTEBOOK_DIR / "..").resolve()
if str(PACKAGE_ROOT) not in sys.path:
    sys.path.insert(0, str(PACKAGE_ROOT))

import annnet as an

from pathlib import Path import sys NOTEBOOK_DIR = Path.cwd() PACKAGE_ROOT = (NOTEBOOK_DIR / "..").resolve() if str(PACKAGE_ROOT) not in sys.path: sys.path.insert(0, str(PACKAGE_ROOT)) import annnet as an

In [2]:

G = an.AnnNet(
    directed=None,
    study="toy stress-signaling network",
    organism="human",
)

G.add_vertices(
    [
        {"vertex_id": "ATM", "kind": "kinase"},
        {"vertex_id": "TP53", "kind": "transcription_factor"},
        {"vertex_id": "MDM2", "kind": "ubiquitin_ligase"},
        {"vertex_id": "CDKN1A", "kind": "effector"},
    ]
)

G.add_edge(
    "ATM",
    "TP53",
    edge_id="atm_activates_tp53",
    edge_directed=True,
    weight=1.0,
    relation="activates",
)

G.add_edge(
    "TP53",
    "MDM2",
    edge_id="tp53_induces_mdm2",
    edge_directed=True,
    weight=0.8,
    relation="induces",
)

G.add_hyperedge(
    head=["TP53", "ATM"],
    tail=["CDKN1A"],
    edge_id="stress_program",
    weight=1.2,
    process="stress_response",
)

G

G = an.AnnNet( directed=None, study="toy stress-signaling network", organism="human", ) G.add_vertices( [ {"vertex_id": "ATM", "kind": "kinase"}, {"vertex_id": "TP53", "kind": "transcription_factor"}, {"vertex_id": "MDM2", "kind": "ubiquitin_ligase"}, {"vertex_id": "CDKN1A", "kind": "effector"}, ] ) G.add_edge( "ATM", "TP53", edge_id="atm_activates_tp53", edge_directed=True, weight=1.0, relation="activates", ) G.add_edge( "TP53", "MDM2", edge_id="tp53_induces_mdm2", edge_directed=True, weight=0.8, relation="induces", ) G.add_hyperedge( head=["TP53", "ATM"], tail=["CDKN1A"], edge_id="stress_program", weight=1.2, process="stress_response", ) G

Out[2]:

<annnet.core.graph.AnnNet at 0x7ca13c7926d0>

In [3]:

print("shape:", G.shape)
print("vertices:", G.vertices())
print("edges:", G.edges())
print("num vertices:", G.num_vertices)
print("num edges:", G.num_edges)
print("graph metadata:", G.uns)

print("shape:", G.shape) print("vertices:", G.vertices()) print("edges:", G.edges()) print("num vertices:", G.num_vertices) print("num edges:", G.num_edges) print("graph metadata:", G.uns)

shape: (4, 3)
vertices: ['ATM', 'TP53', 'MDM2', 'CDKN1A']
edges: ['atm_activates_tp53', 'tp53_induces_mdm2', 'stress_program']
num vertices: 4
num edges: 3
graph metadata: {'study': 'toy stress-signaling network', 'organism': 'human'}

Incidence matrix and AnnData-style accessors¶

AnnNet stores graph structure as a sparse incidence matrix. The obs, var, and uns properties are the closest analogs to the familiar AnnData interface.

In [4]:

X = G.X()
print(type(X).__name__)
print("matrix shape:", X.shape)
X

X = G.X() print(type(X).__name__) print("matrix shape:", X.shape) X

dok_matrix
matrix shape: (8, 8)

Out[4]:

<Dictionary Of Keys sparse matrix of dtype 'float32'
	with 7 stored elements and shape (8, 8)>

In [5]:

G.obs

G.obs

Out[5]:

shape: (4, 2)

vertex_id	kind
str	str
"ATM"	"kinase"
"TP53"	"transcription_factor"
"MDM2"	"ubiquitin_ligase"
"CDKN1A"	"effector"

In [6]:

G.var

G.var

Out[6]:

shape: (3, 3)

edge_id	relation	process
str	str	str
"atm_activates_tp53"	"activates"	null
"tp53_induces_mdm2"	"induces"	null
"stress_program"	null	"stress_response"

Edge surface helpers¶

The graph also exposes convenience helpers that flatten the structure into a more tabular form.

In [7]:

G.edge_list()

G.edge_list()

Out[7]:

[('ATM', 'TP53', 'atm_activates_tp53', 1.0),
 ('TP53', 'MDM2', 'tp53_induces_mdm2', 0.8),
 (None, None, 'stress_program', 1.2)]

In [8]:

G.edges_view()

G.edges_view()

Out[8]:

shape: (3, 13)

edge_id	kind	directed	global_weight	source	target	edge_type	head	tail	members	relation	process	effective_weight
str	str	bool	f64	str	str	str	list[str]	list[str]	list[str]	str	str	f64
"atm_activates_tp53"	"binary"	true	1.0	"ATM"	"TP53"	"regular"	null	null	null	"activates"	null	1.0
"tp53_induces_mdm2"	"binary"	true	0.8	"TP53"	"MDM2"	"regular"	null	null	null	"induces"	null	0.8
"stress_program"	"hyper"	true	1.2	null	null	null	["ATM", "TP53"]	["CDKN1A"]	null	null	"stress_response"	1.2

In [ ]: