Installation failed.

[alhafidzhamdan]

Hi there,

I created a fresh conda environment with conda create -n genewalk python=3.5

and installed genewalk using pip install git+https://github.com/churchmanlab/genewalk.git

but genewalk -hgave me this error:

Traceback (most recent call last):
  File "/exports/igmm/eddie/Glioblastoma-WGS/anaconda/envs/genewalk/bin/genewalk", line 5, in <module>
    from genewalk.cli import main
  File "/exports/igmm/eddie/Glioblastoma-WGS/anaconda/envs/genewalk/lib/python3.5/site-packages/genewalk/cli.py", line 11, in <module>
    from genewalk.nx_mg_assembler import load_network
  File "/exports/igmm/eddie/Glioblastoma-WGS/anaconda/envs/genewalk/lib/python3.5/site-packages/genewalk/nx_mg_assembler.py", line 6, in <module>
    from indra.databases import go_client
  File "/exports/igmm/eddie/Glioblastoma-WGS/anaconda/envs/genewalk/lib/python3.5/site-packages/indra/databases/__init__.py", line 7, in <module>
    from .identifiers import get_identifiers_url, parse_identifiers_url, \
  File "/exports/igmm/eddie/Glioblastoma-WGS/anaconda/envs/genewalk/lib/python3.5/site-packages/indra/databases/identifiers.py", line 302
    if not db_id.startswith(f'{db_ns}{colon}'):
                                            ^
SyntaxError: invalid syntax

Could you help me troubleshoot please?

A

[bgyori]

From the error message, it looks like you are using Python 3.5. GeneWalk works with Python 3.6+ so you'll have to get a newer version of Python.

[alhafidzhamdan]

Thanks- that's running now. FYI i tried random_seed 9999 and that caused the run to stall so I've just removed it and it seems to be running ok.

[ri23]

Thank you @alhafidzhamdan for notifying us about the random seed 9999. Does GeneWalk work for you without issues when inputting another random seed value or is this error specific to value 9999?

[alhafidzhamdan]

Hi @ri23,

I've re-run with seed 1234 and it worked. Then i ran it again with seed 9999 and it again worked. Not sure what i did differently there- it might be something unrelated.

Btw, can you advise me as to how to re-create the gene to Go term network figure? And is there a way to generate multiple genes to a single Go term?

Many thanks in advance. A

[ri23]

Hi @alhafidzhamdan

Thanks for the clarification on the random seeds. Below a python script to recreate the subgraph around 3 chosen genes (Mal, Pllp and Plp1). The script should be relatively easy to adjust if you want choose a GO term (use its go_id from genewalk_results.csv file) and visualize the connected genes instead of the subnetwork for the above chosen genes. See section "Data preprocessing" below.

Best, Robert

#!/usr/bin/env python
# coding: utf-8

# # GeneWalk network visualization

import os
import re
import copy
import pickle as pkl
import pandas as pd
import networkx as nx
import matplotlib.pyplot as plt
plt.rcParams['pdf.fonttype'] = 42

# ### Load GeneWalk multigraph and results

path = '/home/genewalk/qki/'   

filename = 'multi_graph.pkl'
with open(os.path.join(path,filename), 'rb') as f:
    MG = pkl.load(f)

filename = 'genewalk_results.csv'
GW = pd.read_csv(os.path.join(path,filename)) 

# ## QKI subnetwork visualization

# Data preprocessing

# Choose genes of interest
GENES=['MAL','PLLP','PLP1']
labels = {}   
for node in GENES:
    labels[node] = node

#Genes and Neighors
#NB = neighbors
GENES_NB=copy.deepcopy(GENES)
for source in GENES:
    GENES_NB.extend(list(MG.neighbors(source)))
    print(source, len(GENES_NB))
GENES_NB=sorted(list(set(GENES_NB)))

#Subset of Neighbors that are genes
GENES_only_NB=copy.deepcopy(GENES_NB)
for gene in GENES_NB:
    if re.search('GO:',gene):
        GENES_only_NB.remove(gene)

#Enumerate GO annotations of Mal according to GeneWalk ranking
gene = 'MAL'
MAL_GO_NB = list(GW[GW['hgnc_symbol']==gene].sort_values(by='global_padj')['go_id'])
labels_MAL_GO_NB = dict()
for i in range(len(MAL_GO_NB)):
    labels_MAL_GO_NB[MAL_GO_NB[i]] = str(i+1)
MAL_GO_NB_edges = [(gene,gonode) for gonode in MAL_GO_NB]

# ### Generate SubGraph (for plotting)
SG = MG.subgraph(GENES_NB)

SGplot = nx.OrderedGraph() 
SGplot.add_nodes_from(GENES_NB) 
SGplot.add_edges_from((u, v) for (u, v) in SG.edges() if u in SGplot if v in SGplot)

# ### Generate plot
plt.figure(figsize=(4,4))#units: inch

pos = nx.circular_layout(SGplot)

nx.draw(SGplot, pos=pos, node_color='white', with_labels=False, alpha=0.1, node_size=150)
nx.draw_networkx_nodes(SGplot, pos, nodelist=GENES_NB, node_size=150, node_color='#B82225', alpha=1)
nx.draw_networkx_nodes(SGplot, pos, nodelist=GENES_only_NB, node_size=150, node_color='#007EC3', alpha=1)
nx.draw_networkx_edges(SGplot, pos, edgelist=MAL_GO_NB_edges, width=2.0)#draw thick edges for Mal GO annotations

scale_factor = 1.05
posl = copy.deepcopy(pos)
for node in posl:
    posl[node] = scale_factor * pos[node]

#Add labels to the nodes you require
lab = nx.draw_networkx_labels(SGplot, pos=posl, labels=labels, font_size=8,font_weight='bold')
lab = nx.draw_networkx_labels(SGplot, pos=pos, labels=labels_MAL_GO_NB, font_size=10, 
                              font_color='white',font_weight='bold')

filename = 'subnetwork_circular'
plt.savefig(os.path.join(path, filename + '.pdf'),bbox_inches="tight",transparent=True)
plt.savefig(os.path.join(path, filename + '.png'),bbox_inches="tight",transparent=True)

[alhafidzhamdan]

Fantastic thanks!