BioSpecGT

Spectral graph theory for Biological network analysis

• Guarantees high-performance by using Cython
• Friendly interface, easy to use, verbose error messages
• ???
• Profit

Comparison to networkx

• Random generation of a regular graph

Usage

For the full documentation, see the docs.

Graph creation

from BioSpecGT.biospecgt.graph.generator import complete_graph
g = complete_graph(n=100)

Graph algorithms

from BioSpecGT.biospecgt.graph.generator import complete_graph
from BioSpecGT.biospecgt.utils.graphutils import minimum_spanning_tree

g = complete_graph(n=100)
msp = minimum_spanning_tree(g)

Spectral graph theory

from BioSpecGT.biospecgt.linalg.spectral import laplacian_matrix
from BioSpecGT.biospecgt.graph.generator import k_regular_graph

g = k_regular_graph(n=100, k=20, selfloop=False)
l = laplacian_matrix(g)

Spectral clustering

from BioSpecGT.biospecgt.graph.generator import complete_graph
g = complete_graph(n=100)

Importing datasets

from BioSpecGT.biospecgt.graph.generator import complete_graph
g = complete_graph(n=100)

Co-expression networks

from BioSpecGT.biospecgt.bio.coexpression import build_coexpr
from BioSpecGT.biospecgt.bio.base import top10_genes

# gene expression matrix
data = ...

c = build_coexpr(data)
print('Top 10 co-expressed genes:')
print('\n'.join(top10_genes(c)))

Vizualization

from BioSpecGT.biospecgt.graph.generator import complete_graph
g = complete_graph(n=100)