def exportGroupCells(groups,outputdir,dsra2type):
out=[]
for i in range(len(groups)):
icells=groups[i].cells
for j in icells:
ki=[j,"cluster:%s"%(i)] if j not in dsra2type else [j,"cluster:%s"%(i),dsra2type[j]]
out.append(ki)
BioList(out).ex2File("%s/GroupCells.txt"%(outputdir))
def exportSNP(dfsnp,groups,keptMutations,outputdir):
#mutmat[getMutationDistribution(item,dfsnp,groups) for item in keptMutations]
mutmat=[]
for i in keptMutations:
icells=dfsnp[i]
imut=[]
for g in groups:
gcells=g.cells
gmut=[1 if item in icells else 0 for item in gcells]
imut+=gmut
mutmat.append(imut)
#-------------------------------------------
fig,axs=plt.subplots(1,len(groups),sharey=True,gridspec_kw={'width_ratios':[len(item.cells) for item in groups]})
st=0
for i in range(len(groups)):
ed=st+len(groups[i].cells)
axs[i].imshow([item[st:ed] for item in mutmat],cmap="hot",aspect="auto",interpolation="nearest")
st=ed
for ax,l in zip(axs,range(len(groups))):
ax.set_xticklabels([])
ax.set_xlabel(l)
fig.add_subplot(111,frameon=False)
plt.tick_params(labelcolor='none',top='off',bottom='off',left='off',right='off')
plt.grid(False)
plt.ylabel("SNPs")
plt.xlabel("Clusters")
plt.savefig("%s/SNP_matrix.png"%(outputdir),dpi=600)
mutmatout=[]
headers=['SNP']+reduce(lambda x,y:x+y,[item.cells for item in groups])
mutmatout.append(headers)
for i in range(len(keptMutations)):
mutmat[i]=[keptMutations[i]]+mutmat[i]
mutmatout.append(mutmat[i])
BioList(mutmatout).ex2File("%s/SNP_matrix.tsv"%(outputdir))
def drawTree(net,outstr,outputdir):
pos=networkx.spring_layout(net)
#Xlabels={net.nodes.keys()[k]:net.nodes.keys()[k].name.split("|")[0] for k in range(len(net.nodes))}
Xlabels={}
for k in range(len(net.nodes)):
nkeys=list(net.nodes.keys())
Xlabels[nkeys[k]]=nkeys[k].name.split("|")[0]
XlabelTexts=[item.name.split("|") for item in Xlabels]
Xnames=[item[0] for item in XlabelTexts]
XlabelTexts=[item[-1] if len(item)>1 else "" for item in XlabelTexts]
networkx.draw(net,pos,labels=Xlabels,with_labels=True)
datgraph=[]
for i in range(len(pos.values())):
[x,y]=list(pos.values())[i]
datgraph.append([Xnames[i],(x,y)])
plt.text(x-0.003*len(XlabelTexts[i]),y+0.05,s=XlabelTexts[i],fontsize=5)
plt.subplots_adjust(right=0.5)
plt.savefig("%s/%s.png"%(outputdir,outstr),bbox_inches="tight",dpi=600)
BioList(datgraph).ex2File("%s/%s.dat"%(outputdir,outstr))
def main():
parser = argparse.ArgumentParser()
parser.add_argument(
"-i",
"--input_genes",
required=True,
help="the input gene list for the SARS-CoV-2 association analysis")
parser.add_argument(
"-v",
"--viral_host_interactions",
required=True,
help="The interactions between viral proteins and host proteins")
parser.add_argument("-p",
"--host_protein_interactions",
required=True,
help="The interaction between host proteins")
parser.add_argument("-o",
"--output",
required=True,
help="The specified output directory")
parser.add_argument(
"-b",
"--background",
required=False,
default=1000,
help=
"Integer, the number of background proteins/genes used to calculate the connectivity significance of input genes"
)
args = parser.parse_args()
fnCOVID19 = args.viral_host_interactions
fnPPI = args.host_protein_interactions
fnMarker = args.input_genes
output = args.output
try:
bgN = int(args.background)
except:
print("-b must be integer, please check your input!")
sys.exit(0)
#----------------------------
# read in covid19 interactions
print("processing input files ...")
COVInteractions = TabFile(fnCOVID19).read("\t")[1:]
SProteins = list(set([item[1] for item in COVInteractions]))
source = list(set([item.split(" ")[0] for item in SProteins]))
#pdb.set_trace()
dCI = {}
for i in COVInteractions:
dCI[i[1].split(" ")[0] + "," + i[3]] = [i[1], i[3]]
COVInteractions = [[item[1].split(" ")[0], item[3], item[-2]]
for item in COVInteractions]
COVProteins = []
for i in COVInteractions:
COVProteins += i[:2]
COVProteins = list(set(COVProteins))
# read in PPIs
PPI = TabFile(fnPPI).read("\t")[1:]
PPI = [[item[0], item[2], item[3]] for item in PPI]
# read in markers
markers = LineFile(fnMarker).read()
markers = [item for item in markers if item != '']
print("building protein networks ...")
AllN = getNodes(COVInteractions) + getNodes(PPI) + markers
AllNodes = []
for i in AllN:
if i not in AllNodes:
AllNodes.append(i)
#AllNodes=list(set(AllN))
AEdges = getEdges(COVInteractions + PPI)
AllEdges = {}
for i in AEdges:
iKey = i
iScore = AEdges[i]
if iScore > 0:
iScore = math.log(1.0 / iScore, 2)
AllEdges[iKey] = iScore
G = nx.Graph()
for i in AllNodes:
G.add_node(i)
for i in AllEdges:
[A, B] = i.split(',')
si = AllEdges[i]
G.add_edge(A, B, weight=si)
print("learning optimal path from SARS-CoV-2 to input genes ...")
targets = markers
SPaths = []
AllProteins = []
AllUProteins = []
for i in source:
for j in targets:
if nx.has_path(G, i, j):
pij_all = nx.all_shortest_paths(G, i, j)
sij_all = []
for pijk in pij_all:
sijk = getScore(G, pijk)
sij_all.append([sijk, pijk])
sij_all = sorted(sij_all, key=lambda x: x[0])
[sij, pij] = sij_all[0]
AllProteins += pij
uij = dCI[pij[0] + ',' + pij[1]] + pij[2:]
AllUProteins += uij
SPaths.append([i, j, sij, uij])
print(j)
AllProteins = list(set(AllProteins))
outEdges = extractEdges(AllProteins, AEdges)
network = []
for i in outEdges:
[A, B] = i.split(",")
iScore = outEdges[i]
if i in dCI:
[A, B] = dCI[i]
ni = [A, iScore, B]
network.append(ni)
NodeInfo = [['Node', 'Type']]
for i in AllUProteins:
if i in SProteins:
NodeInfo.append([i, 'virus'])
elif i in COVProteins:
NodeInfo.append([i, 'source'])
elif i in targets:
NodeInfo.append([i, 'target'])
else:
NodeInfo.append([i, 'Intermediate'])
print(
"finding the optimal paths from SARS-CoV-2 proteins to all host proteins"
)
APaths = []
ct = 0
# fix the random initialization
random.seed(a=10)
bgNodes = random.sample(AllNodes[1:], bgN)
for i in source:
for j in bgNodes:
if nx.has_path(G, i, j):
pij_all = nx.all_shortest_paths(G, i, j)
sij_all = []
for pijk in pij_all:
sijk = getScore(G, pijk)
sij_all.append([sijk, pijk])
sij_all = sorted(sij_all, key=lambda x: x[0])
[sij, pij] = sij_all[0]
APaths.append([i, j, sij, pij])
ct += 1
print(ct)
SPath = [float(item[2]) for item in SPaths]
APath = [float(item[2]) for item in APaths]
pv1 = mannwhitneyu(SPath, APath, alternative='less')
print("p-value: %s" % (pv1[1]))
XX = [SPath, APath]
#df=pd.DataFrame(data=XX)
#df.index=['HLH Genes','All Genes']
if os.path.exists(output) == False:
os.mkdir(output)
print("writing results ...")
print("exporting inferred paths ...")
BioList(SPaths).ex2File("%s/SPaths.txt" % (output), "\t")
BioList(APaths).ex2File("%s/APaths.txt" % (output), "\t")
print("exporting network file (.sif) ...")
BioList(network).ex2File("%s/network.sif" % (output), '\t')
print("exporting network node attribute file (.txt)")
BioList(NodeInfo).ex2File("%s/NodeInfo.txt" % (output), '\t')
sns.boxplot(data=XX)
plt.xticks(range(len(XX)), ['Input Genes', 'All Genes'])
plt.ylabel("Connectivity Score")
plt.savefig("%s/Connectivity.pdf" % (output))