def peaks_only(config_file, untagged, organism):
CP_out = []
quant_bams = {}
with open(config_file, 'r') as config:
for line in config:
if untagged in line:
CP_untagged = line.strip()
elif 'changepoint' in line.lower() or 'peak' in line.lower():
CP_out.append(line.strip())
#bam files for quantitation should be file,quant,A1
elif 'quant' in line:
quant_bams[line.split(',')[-1].strip()] = line.split(',')[0]
name = config_file.split('/')[-1].split('_config')[0]
base_dir = config_file.split(name)[0]
if base_dir == '': base_dir = './'
print "Output file location and prefix: "+base_dir+name
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
peak_df = SP.peak_to_seq_pipeline(CP_untagged, CP_out[0], CP_out[1], gff3, fa_dict, name=name+'_CP_peaks')
peak_df.to_pickle(base_dir+name+'_all_peaks.pickle')
quant_df = SP.quant_from_peak_df(peak_df, gff3, fa_dict, organism=organism)
quant_df = SP.quantitate_junction_df(quant_bams, quant_df, gff3, organism=organism)
quant_df.to_pickle(base_dir+name+'_quantitation.pickle')
quant_df.to_csv(base_dir+name+'_quantitation.csv')
scatter = SP.SP_pipeline_scatters(quant_df, base_dir, name)
def position_wise_scores2(seq5_list, seq3_list, organism, title='Intron position strength'):
'''Uses chi-contingency test to score base proportions at each position in sample against population'''
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
all_5p, all_3p = generate_all_ss_seqs(gff3, fa_dict, organism)
pop_5p = seq_list_to_totals(all_5p)
pop_3p = seq_list_to_totals(all_3p)
samp_5p = seq_list_to_totals(seq5_list)
samp_3p = seq_list_to_totals(seq3_list)
print samp_5p.shape
p5 = []
for n in range(samp_5p.shape[1]):
if n == 2 or n == 3:
p5.append(1)
else:
conting = np.array([samp_5p[:,n],pop_5p[:,n]])
chi2, p, dof, expected = stats.chi2_contingency(conting)
p5.append(np.log10(p)*-1)
p3 = []
for n in range(samp_3p.shape[1]):
if n == 4 or n == 5:
p3.append(1)
else:
conting = np.array([samp_3p[:,n],pop_3p[:,n]])
chi2, p, dof, expected = stats.chi2_contingency(conting)
p3.append(np.log10(p)*-1)
fig, ax = plt.subplots(2, 1, figsize=(4,4))
width = 0.7
max_y = max(p5+p3) + 0.1*max(p5+p3)
ind5 = np.arange(len(p5))
ax[0].bar(ind5, p5, color='k')
ax[0].plot([0,8], [2,2], '--', color='0.7')
ax[0].set_xlim([0,len(p5)])
ax[0].set_ylabel("5' splice site\n-log10(p-value)")
ax[0].set_title(title)
ax[0].set_ylim([0,max_y])
ind3 = np.arange(len(p3))
ax[1].bar(ind3, p3, color='k')
ax[1].plot([0,8], [2,2], '--', color='0.7')
ax[1].set_xlim([0,len(p3)])
ax[1].set_ylabel("3' splice site\n-log10(p-value)")
ax[1].set_ylim([0,max_y])
ax[0].set_xticks(ind3 + width / 2)
ax[1].set_xticks(ind3 + width / 2)
ax[0].set_xticklabels(np.arange(-2,6))
ax[1].set_xticklabels(np.arange(-5,3))
fig.tight_layout()
plt.show()
return fig
def peak_seq_enrichment(df, organism):
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
nuc_prob = SP.gc_content(fa_dict)
p_dict = {'A':nuc_prob[0], 'T':nuc_prob[2], 'C':nuc_prob[1], 'G':nuc_prob[3]}
unpeaks = df[df['type'] == 'other']
unpeaks = unpeaks.append(df[df['type'] == 'intronic'])
print "Number of unpredicted peaks:"
print len(unpeaks)
nucs = ['G','A','C','T']
dinucs = set()
for nuc in nucs:
for nuc2 in nucs:
dinucs.add(nuc+nuc2)
five = {}
three = {}
for dinuc in dinucs:
five[dinuc] = len(unpeaks[unpeaks['sequence'].str[6:8].str.contains(dinuc)])
three[dinuc] = len(unpeaks[unpeaks['sequence'].str[4:6].str.contains(dinuc)])
five_LO = {}
three_LO = {}
for dinuc in five.keys():
p_dinuc = p_dict[dinuc[0]]*p_dict[dinuc[1]]
phat_dinuc = five[dinuc]/float(len(unpeaks))
phat_dinuc2 = three[dinuc]/float(len(unpeaks))
SE = np.sqrt(phat_dinuc*(1-phat_dinuc)/len(unpeaks))
SE2 = np.sqrt(phat_dinuc2*(1-phat_dinuc2)/len(unpeaks))
Z = (phat_dinuc-p_dinuc)/SE
Z2 = (phat_dinuc2-p_dinuc)/SE2
pvalue = stats.norm.sf(Z)
pvalue2 = stats.norm.sf(Z2)
LO = np.log((1-pvalue)/pvalue)
LO2 = np.log((1-pvalue2)/pvalue2)
five_LO[dinuc] = LO
three_LO[dinuc] = LO2
fig, ax = plt.subplots(figsize=(12,6))
width = 0.35
ind = np.arange(len(five_LO.keys()))
rects2 = ax.bar(ind, three_LO.values(), width, color='crimson', edgecolor='crimson', label='Before peak')
rects1 = ax.bar(ind + width, five_LO.values(), width, color='indigo', edgecolor='indigo', label='After peak')
ax.plot([-1,17],[0,0],'-', color='black')
ax.plot([-1,17],[2.94,2.94], '--', color='0.7', label='95% CI')
ax.plot([-1,17],[-2.94,-2.94], '--', color='0.7')
ax.set_xlim([-1,17])
ax.set_xticklabels(five_LO.keys(), fontsize=12)
ax.set_xticks(ind + width / 2)
ax.set_ylabel('Log odds dinucleotide enrichment', fontsize=14)
ax.set_title('Unpredicted peaks', fontsize=14)
ax.legend(fontsize=12)
return fig
def peaks_only(config_file, untagged, organism):
CP_out = []
quant_bams = {}
with open(config_file, 'r') as config:
for line in config:
if untagged in line:
CP_untagged = line.strip()
elif 'changepoint' in line.lower() or 'peak' in line.lower():
CP_out.append(line.strip())
#bam files for quantitation should be file,quant,A1
elif 'quant' in line:
quant_bams[line.split(',')[-1].strip()] = line.split(',')[0]
name = config_file.split('/')[-1].split('_config')[0]
base_dir = config_file.split(name)[0]
if base_dir == '': base_dir = './'
print "Output file location and prefix: " + base_dir + name
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
peak_df = SP.peak_to_seq_pipeline(CP_untagged,
CP_out[0],
CP_out[1],
gff3,
fa_dict,
name=name + '_CP_peaks')
peak_df.to_pickle(base_dir + name + '_all_peaks.pickle')
quant_df = SP.quant_from_peak_df(peak_df, gff3, fa_dict, organism=organism)
quant_df = SP.quantitate_junction_df(quant_bams,
quant_df,
gff3,
organism=organism)
quant_df.to_pickle(base_dir + name + '_quantitation.pickle')
quant_df.to_csv(base_dir + name + '_quantitation.csv')
scatter = SP.SP_pipeline_scatters(quant_df, base_dir, name)
def position_wise_scores(seq_5p, seq_3p, organism):
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
PSSM_5p, PSSM_3p = generate_consensus_matrix(gff3, fa_dict, PSSM=True)
base_dict = {"A":0, "C":1, "T":2, "G":3}
seq_5p = [x for x in seq_5p if x is not None]
seq_3p = [x for x in seq_3p if x is not None]
score_5prime = np.empty([2,len(seq_5p[0])])
score_3prime = np.empty([2,len(seq_3p[0])])
all_5p = np.empty([len(seq_5p), len(seq_5p[0])])
all_3p = np.empty([len(seq_3p), len(seq_3p[0])])
n=0
for n in range(len(seq_5p)):
for a, base in enumerate(seq_5p[n]):
all_5p[n,a] = PSSM_5p[base_dict[base], a]
a=0
for a in range(len(score_5prime[0])):
score_5prime[0,a] = np.median(all_5p[0:,a])
score_5prime[1,a] = (max(all_5p[0:,a])-min(all_5p[0:,a]))/2.
print score_5prime
m=0
for m in range(len(seq_3p)):
for b, base in enumerate(seq_3p[m]):
all_3p[m,b] = PSSM_3p[base_dict[base], b]
b=0
for b in range(len(score_3prime[0])):
score_3prime[0,b] = np.median(all_3p[0:,b])
score_3prime[1,b] = (max(all_3p[0:,b])-min(all_3p[0:,b]))/2.
print score_3prime
return all_5p, all_3p
def position_wise_scores(seq_5p, seq_3p, organism):
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
PSSM_5p, PSSM_3p = generate_consensus_matrix(gff3, fa_dict, PSSM=True)
base_dict = {"A": 0, "C": 1, "T": 2, "G": 3}
seq_5p = [x for x in seq_5p if x is not None]
seq_3p = [x for x in seq_3p if x is not None]
score_5prime = np.empty([2, len(seq_5p[0])])
score_3prime = np.empty([2, len(seq_3p[0])])
all_5p = np.empty([len(seq_5p), len(seq_5p[0])])
all_3p = np.empty([len(seq_3p), len(seq_3p[0])])
n = 0
for n in range(len(seq_5p)):
for a, base in enumerate(seq_5p[n]):
all_5p[n, a] = PSSM_5p[base_dict[base], a]
a = 0
for a in range(len(score_5prime[0])):
score_5prime[0, a] = np.median(all_5p[0:, a])
score_5prime[1, a] = (max(all_5p[0:, a]) - min(all_5p[0:, a])) / 2.
print score_5prime
m = 0
for m in range(len(seq_3p)):
for b, base in enumerate(seq_3p[m]):
all_3p[m, b] = PSSM_3p[base_dict[base], b]
b = 0
for b in range(len(score_3prime[0])):
score_3prime[0, b] = np.median(all_3p[0:, b])
score_3prime[1, b] = (max(all_3p[0:, b]) - min(all_3p[0:, b])) / 2.
print score_3prime
return all_5p, all_3p
def main():
'''Each line will be : bam_file,genotype,sample
e.g. CM763-A_sorted.bam,WT,A1'''
bam_dict = {}
with open(sys.argv[1], 'r') as config:
for line in config:
info = line.split(',')
genotype = info[1]
sample = info[2].strip()
if genotype not in bam_dict:
bam_dict[genotype] = {}
bam_dict[genotype][sample] = info[0]
prefix = sys.argv[1].split('_config')[0]
organism = sys.argv[3]
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
columns = ['5p score','exon size (us)','exon size (ds)','introns in transcript','type','transcript size','intron size',
'chromosome','position','alt splicing','3p score','transcript','intron position','strand','peak',
'Base 5-0','Base 5-1','Base 5-2','Base 5-3','Base 5-4','Base 5-5','Base 5-6','Base 5-7','Base 3-0',
'Base 3-1','Base 3-2','Base 3-3','Base 3-4','Base 3-5','Base 3-6','Base 3-7','branch score',
'branch to 3p distance','percent pPy','branch-0','branch-1','branch-2','branch-3','branch-4']
quant_df = pd.read_csv(sys.argv[2], index_col=0)
try:
quant_df = quant_df[columns]
except KeyError:
print "Columns missing from dataframe..."
print columns
print quant_df.columns
return None
final_df = copy.deepcopy(quant_df)
final_df.columns = pd.MultiIndex.from_product([['Peaks'], final_df.columns])
for genotype, samples in bam_dict.iteritems():
# Determine if whole cell extract samples are present
Ws = [x for x in samples.keys() if "W" in x]
if len(Ws) > 1:
W = True
else:
W=False
# Quantitate all samples with genotype
new_df = quantitate_junction_df(samples, quant_df, gff3, W=W)
# Remove original columns and rename new ones with multiindex
new_columns = [x for x in new_df.columns if x not in columns]
new_df = new_df[new_columns]
new_df.columns = pd.MultiIndex.from_product([[genotype], new_df.columns])
final_df = final_df.join(new_df, how='inner')
#final_df = final_df.merge(new_df, right_index=True, left_index=True)
final_df.to_csv(prefix+'_quant_df.csv')
final_df.to_pickle(prefix+'_quant_df.pickle')
SP.SP_quant_scatters(final_df.dropna(how='any'), bam_dict, W=W)
def main():
'''Each line will be : bam_file,genotype,sample
e.g. CM763-A_sorted.bam,WT,A1'''
bam_dict = {}
with open(sys.argv[1], 'r') as config:
for line in config:
info = line.split(',')
genotype = info[1]
sample = info[2].strip()
if genotype not in bam_dict:
bam_dict[genotype] = {}
bam_dict[genotype][sample] = info[0]
prefix = sys.argv[1].split('_config')[0]
organism = sys.argv[3]
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
columns = [
'5p score', 'exon size (us)', 'exon size (ds)',
'introns in transcript', 'type', 'transcript size', 'intron size',
'chromosome', 'position', 'alt splicing', '3p score', 'transcript',
'intron position', 'strand', 'peak', 'Base 5-0', 'Base 5-1',
'Base 5-2', 'Base 5-3', 'Base 5-4', 'Base 5-5', 'Base 5-6', 'Base 5-7',
'Base 3-0', 'Base 3-1', 'Base 3-2', 'Base 3-3', 'Base 3-4', 'Base 3-5',
'Base 3-6', 'Base 3-7', 'branch score', 'branch to 3p distance',
'percent pPy', 'branch-0', 'branch-1', 'branch-2', 'branch-3',
'branch-4'
]
quant_df = pd.read_csv(sys.argv[2], index_col=0)
try:
quant_df = quant_df[columns]
except KeyError:
print "Columns missing from dataframe..."
print columns
print quant_df.columns
return None
final_df = copy.deepcopy(quant_df)
final_df.columns = pd.MultiIndex.from_product([['Peaks'],
final_df.columns])
for genotype, samples in bam_dict.iteritems():
# Determine if whole cell extract samples are present
Ws = [x for x in samples.keys() if "W" in x]
if len(Ws) > 1:
W = True
else:
W = False
# Quantitate all samples with genotype
new_df = quantitate_junction_df(samples, quant_df, gff3, W=W)
# Remove original columns and rename new ones with multiindex
new_columns = [x for x in new_df.columns if x not in columns]
new_df = new_df[new_columns]
new_df.columns = pd.MultiIndex.from_product([[genotype],
new_df.columns])
final_df = final_df.join(new_df, how='inner')
#final_df = final_df.merge(new_df, right_index=True, left_index=True)
final_df.to_csv(prefix + '_quant_df.csv')
final_df.to_pickle(prefix + '_quant_df.pickle')
SP.SP_quant_scatters(final_df.dropna(how='any'), bam_dict, W=W)
def main():
'''Usage: run SP_pipeline.py config_file untagged_sample_name organism
config file : file that lists all branch, junction and peak files
untagged_sample_name : prefix for untagged sample
organism : pombe, crypto or cerevisiae'''
junc_beds = []
branch_bams = []
CP_out = []
CP_untagged = None
quant_bams = {}
# Read configuration file
with open(sys.argv[1], 'r') as config:
for line in config:
if 'junctions.bed' in line.lower():
junc_beds.append(line.strip())
elif 'branch' in line.lower():
branch_bams.append(line.strip())
elif sys.argv[2] in line:
CP_untagged = line.strip()
elif 'changepoint' in line.lower() or 'peak' in line.lower():
CP_out.append(line.strip())
#bam files for quantitation should be file,quant,A1
elif 'quant' in line:
quant_bams[line.split(',')[-1].strip()] = line.split(',')[0]
name = sys.argv[1].split('/')[-1].split('_config')[0]
base_dir = sys.argv[1].split(name)[0]
if base_dir == '': base_dir = './'
print "Output file location and prefix: " + base_dir + name
print "\nJunction bed files"
print junc_beds
print "\nBranch bam files"
if len(branch_bams) == 2:
print branch_bams
use_branches = True
elif len(branch_bams) == 0:
print "No data for branches, continuing with only junctions"
use_branches = False
print "\nUntagged peaks"
print CP_untagged
print "\nChangepoint peaks"
print CP_out
print ''
if CP_untagged is None:
print "\n Error: no untagged file indicated"
return None
organism = sys.argv[3]
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
#### Generate peak df
if name + '_peaks_w_branch.csv' not in os.listdir(
base_dir) or name + '_peaks_w_junc.csv' not in os.listdir(
base_dir):
if name + '_all_peaks.pickle' not in os.listdir(base_dir):
peak_df = SP.peak_to_seq_pipeline(CP_untagged,
CP_out[0],
CP_out[1],
gff3,
fa_dict,
name=name + '_CP_peaks')
peak_df.to_pickle(base_dir + name + '_all_peaks.pickle')
else:
peak_df = pd.read_pickle(base_dir + name + '_all_peaks.pickle')
#### Junction to peak comparison
if name + '_peaks_w_junc.csv' not in os.listdir(base_dir):
print "Generating peaks vs. junctions dataframe..."
peaks_w_junc = peak_junction_analysis(peak_df, junc_beds, gff3,
fa_dict, organism, base_dir,
name)
else:
peaks_w_junc = pd.read_pickle(base_dir + name + '_peaks_w_junc.pickle')
print "Peaks vs. junction dataframe already exists"
#### Branch to peak comparison
if use_branches is True:
if name + '_peaks_w_branch.csv' not in os.listdir(base_dir):
print "Generating peaks vs. branches dataframe..."
peaks_w_branch = peak_branch_analysis(peak_df, branch_bams, gff3,
fa_dict, organism, base_dir,
name)
else:
peaks_w_branch = pd.read_pickle(base_dir + name +
'_peaks_w_branch.pickle')
print "Peaks vs. branches dataframe already exists"
#### Clean up dataframe for quantitation
if name + '_quantitation.csv' not in os.listdir(base_dir):
quant_df, lariat_df = SP.make_quant_df(peaks_w_junc,
peaks_w_branch,
gff3,
fa_dict,
organism=organism)
quant_df = SP.find_score_branches_ppy(quant_df, peaks_w_branch,
fa_dict)
print "Counting reads in transcripts and at peaks..."
quant_df = SP.quantitate_junction_df(quant_bams,
quant_df,
gff3,
organism=organism)
quant_df.to_pickle(base_dir + name + '_quantitation.pickle')
quant_df.to_csv(base_dir + name + '_quantitation.csv')
lariat_df.to_pickle(base_dir + name + '_lariats.pickle')
lariat_df.to_csv(base_dir + name + '_lariats.csv')
scatter = SP.SP_pipeline_scatters(quant_df, base_dir, name)
else:
quant_df = pd.read_pickle(base_dir + name + '_quantitation.pickle')
scatter = SP.SP_pipeline_scatters(quant_df, base_dir, name)
print "\n****Finished****"
def igv_plots_general(bam_list,
gene_list,
organism,
colors=None,
names=None,
save_dir=None,
unstranded=False,
end_only=False,
same_yaxis=False,
specific_range=None,
transcript_direction=True,
log_scale=False,
rpm=True,
PE=False,
plot_junctions=False):
'''Usage:
Parameters
----------
bam_list : list, bam files in order of plotting (top to bottom)
gene_list : list of transcripts to plot (should be genes not transcript isoforms)
if dataframe passed instead of list, will plot introns (must have intron information in datafame)
organism : str, pombe or crypto
colors : list, default `None`
list of colors to use, same length as bam_list, check matplotlib documentation for valid color names
names : list, default `None`
list of sample names to use instead of bam file names. Same length as bam_files
save_dir : str, default `None`
directory to save eps files. If None, does not save files
unstranded : bool, default `False`
Use True for ChIP or DNA sequencing data (or unstranded RNAseq)
end_only : bool or list, default `False`
Whether to plot only the ends of reads. If different for each bam, make a list of bools same length as bam_list
same_yaxis : bool, default `False`
Whether all samples should be plotted on the same axis after normalizing to total number of aligned reads
specific_range : str, default `None`
Options: ('end', window)
('start', window)
([coordinate], window)
transcript_direction : bool, default `True`
If True, will plot in the direction of transcription, not in the direction of the DNA
'''
# Get all organism information (annotation etc.)
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
tx_dict = SP.build_transcript_dict(gff3, organism=organism)
fix_info = {
'I': 'chr1',
'II': 'chr2',
'III': 'chr3',
'chr1': 'I',
'chr2': 'II',
'chr4': 'IV',
'chr5': 'V',
'chr6': 'VI',
'chr7': 'VII',
'chr8': 'VIII',
'chr9': 'IX',
'chr10': 'X',
'chr11': 'XI',
'chr12': 'XII',
'chr13': 'XIII',
'chr14': 'XIV',
'chr15': 'XV',
'chr16': 'XVI',
'-': '+',
'+': '-',
'chr1': 'I',
'chr2': 'II',
'chr3': 'III'
}
if organism == 'pombe':
tx_suffix = '.1'
else:
tx_suffix = 'T0'
# Set up range parameters if specific range is indicated
if specific_range is not None:
window = int(specific_range[1])
new_tx_dict = {}
for gene in gene_list:
info = tx_dict[gene + tx_suffix]
if specific_range[0] == 'end':
if info[2] == '+':
start = info[1] - window
end = info[1] + window
else:
start = info[0] - window
end = info[0] + window
elif specific_range[0] == 'start':
if info[2] == '-':
start = info[1] - window
end = info[1] + window
else:
start = info[0] - window
end = info[0] + window
else:
start = int(specific_range[0]) - window
end = int(specific_range[0]) + window
new_tx_dict[gene + tx_suffix] = [start, end, info[2], info[3]]
else:
new_tx_dict = tx_dict
# Open bam files and count reads if rpm is True
open_bams = {}
total_list = []
for bam in bam_list:
open_bams[bam] = pysam.Samfile(bam)
if rpm is True:
total = check_output(['samtools', 'view', '-F 0x04', '-c',
bam]).strip()
total = float(total) / 1000000.
total_list.append(total)
else:
total_list.append(1.)
# Expand optional arguments to lists if necessary
colors = list_from_arg(colors, len(bam_list))
end_only = list_from_arg(end_only, len(bam_list))
log_scale = list_from_arg(log_scale, len(bam_list))
unstranded = list_from_arg(unstranded, len(bam_list))
# Get gene_list from dataframe if gene_list is not a list
df = None
if type(gene_list) == dict:
new_tx_dict = gene_list
gene_list = gene_list.keys()
elif type(gene_list) != list:
df = gene_list
gene_list = df.index
for tx in gene_list:
num_ax = len(bam_list) + 1
if plot_junctions is True:
num_ax += len(bam_list)
fig, ax = plt.subplots(num_ax, figsize=(10, num_ax), sharex=True)
fig.subplots_adjust(hspace=0)
# Get transcript info from transcript_dictionary
if df is None:
try:
info = new_tx_dict[tx + tx_suffix]
except KeyError:
info = new_tx_dict[tx]
chrom = info[3]
start = info[0]
end = info[1]
strand = info[2]
# If dataframe was passed, get plotting information from dataframe instead
else:
if isinstance(df.columns, pd.core.index.MultiIndex):
new_columns = [x[1] for x in df.columns if x[0] == 'Peaks']
df = df[[x for x in df.columns if x[0] == 'Peaks']]
df.columns = new_columns
strand = df.loc[tx, 'strand']
chrom = df.loc[tx, 'chromosome']
if strand == '+':
start = df.loc[tx, 'position'] - 100
end = df.loc[tx, 'position'] + df.loc[tx, 'intron size'] + 100
elif strand == '-':
start = df.loc[tx, 'position'] - df.loc[tx,
'intron size'] - 100
end = df.loc[tx, 'position'] + 100
start = int(start)
end = int(end)
tx = df.loc[tx, 'transcript']
# Generate read series for each transcript
max_y = 0
junc_ymax = 0
for n, bam in enumerate(bam_list):
try:
bam_iter = open_bams[bam].fetch(chrom, start, end)
except ValueError:
chrom = fix_info[chrom]
bam_iter = open_bams[bam].fetch(chrom, start, end)
if end_only[n] is not False:
s = SP.generate_read_series_A(bam_iter, chrom, start, end,
strand)
linewidth = 2
else:
if PE is False:
s = SP.generate_read_series_B(bam_iter, chrom, start, end,
strand)
else:
s = SP.generate_read_series_PE(bam_iter, chrom, start, end,
strand)
linewidth = 1
# Get reads from otherstrand if the library type is unstranded
if unstranded[n] is True:
bam_iter = open_bams[bam].fetch(chrom, start, end)
if end_only[n] is not False:
s2 = SP.generate_read_series_A(bam_iter, chrom, start, end,
fix_info[strand])
linewidth = 2
else:
if PE is False:
s2 = SP.generate_read_series_B(bam_iter, chrom, start,
end, fix_info[strand])
else:
s2 = SP.generate_read_series_PE(
bam_iter, chrom, start, end, fix_info[strand])
linewidth = 1
s = s.add(s2)
# Normalize to rpm (will just divide by 1 if rpm is False)
s = s.divide(total_list[n])
if log_scale[n] is True:
s = s.apply(np.log2)
# Plot!
ax[n].bar(s.index,
s,
linewidth=linewidth,
color=colors[n],
edgecolor=colors[n],
zorder=2)
ax[n].tick_params(axis='both', which='major', labelsize=14)
max_y = max([max_y, max(s)])
if plot_junctions is True:
m = n + len(bam_list)
intron_dict = get_junctions(open_bams[bam], chrom, start, end,
strand)
ax[m].plot((start, end), (0, 0), '-', c='k')
for coords, heights in intron_dict.iteritems():
ax[m].plot(coords,
heights,
'-',
linewidth=2,
color=colors[n])
ax[m].fill_between(coords,
0,
heights,
facecolor=colors[n],
interpolate=True,
alpha=0.5)
if same_yaxis is True:
junc_ymax = max(
[junc_ymax,
max(zip(*intron_dict.values())[1])])
# Add diagram of gene below traces
if tx in tx_dict:
strand = gene_patches(tx, tx_dict, ax[-1])
ax[-1].set_xlim(start, end)
else:
try:
new_tx = tx.split(' ')[0]
if new_tx[-2] == 'T' or new_tx[-2] == '.':
new_tx = new_tx[:-2]
strand = gene_patches(new_tx, tx_dict, ax[-1])
ax[-1].set_xlim(start, end)
except KeyError:
print "Transcript unknown"
# Flip minus strand transcripts if indicated
if transcript_direction is True:
if strand == '-':
ax[-1].invert_xaxis()
# Set x and y limits
for n in range(len(bam_list)):
ax[n].set_xlim(start, end)
if same_yaxis is True:
ax[n].set_ylim(0, max_y + 0.1 * max_y)
if plot_junctions is True:
ax[n + len(bam_list)].set_ylim(0,
junc_ymax + 0.1 * junc_ymax)
if strand == '-':
ax[n].invert_xaxis()
ax[0].set_ylabel('RPM', fontsize=16)
ax[0].set_title(tx, fontsize=16)
#ax[0].get_xaxis().set_ticks([])
plt.show()
# Save if indicated
if save_dir is not None:
if not os.path.exists(save_dir):
os.makedirs(save_dir)
fig.savefig(save_dir + tx + '.eps', format='eps')
plt.clf()
def main():
'''Usage: run SP_pipeline.py config_file untagged_sample_name organism
config file : file that lists all branch, junction and peak files
untagged_sample_name : prefix for untagged sample
organism : pombe, crypto or cerevisiae'''
junc_beds = []
branch_bams = []
CP_out = []
CP_untagged = None
quant_bams = {}
# Read configuration file
with open(sys.argv[1], 'r') as config:
for line in config:
if 'junctions.bed' in line.lower():
junc_beds.append(line.strip())
elif 'branch' in line.lower():
branch_bams.append(line.strip())
elif sys.argv[2] in line:
CP_untagged = line.strip()
elif 'changepoint' in line.lower() or 'peak' in line.lower():
CP_out.append(line.strip())
#bam files for quantitation should be file,quant,A1
elif 'quant' in line:
quant_bams[line.split(',')[-1].strip()] = line.split(',')[0]
name = sys.argv[1].split('/')[-1].split('_config')[0]
base_dir = sys.argv[1].split(name)[0]
if base_dir == '': base_dir = './'
print "Output file location and prefix: "+base_dir+name
print "\nJunction bed files"
print junc_beds
print "\nBranch bam files"
if len(branch_bams) == 2:
print branch_bams
use_branches = True
elif len(branch_bams) == 0:
print "No data for branches, continuing with only junctions"
use_branches = False
print "\nUntagged peaks"
print CP_untagged
print "\nChangepoint peaks"
print CP_out
print ''
if CP_untagged is None:
print "\n Error: no untagged file indicated"
return None
organism = sys.argv[3]
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
#### Generate peak df
if name+'_peaks_w_branch.csv' not in os.listdir(base_dir) or name+'_peaks_w_junc.csv' not in os.listdir(base_dir):
if name+'_all_peaks.pickle' not in os.listdir(base_dir):
peak_df = SP.peak_to_seq_pipeline(CP_untagged, CP_out[0], CP_out[1], gff3, fa_dict, name=name+'_CP_peaks')
peak_df.to_pickle(base_dir+name+'_all_peaks.pickle')
else:
peak_df = pd.read_pickle(base_dir+name+'_all_peaks.pickle')
#### Junction to peak comparison
if name+'_peaks_w_junc.csv' not in os.listdir(base_dir):
print "Generating peaks vs. junctions dataframe..."
peaks_w_junc = peak_junction_analysis(peak_df, junc_beds, gff3, fa_dict, organism, base_dir, name)
else:
peaks_w_junc = pd.read_pickle(base_dir+name+'_peaks_w_junc.pickle')
print "Peaks vs. junction dataframe already exists"
#### Branch to peak comparison
if use_branches is True:
if name+'_peaks_w_branch.csv' not in os.listdir(base_dir):
print "Generating peaks vs. branches dataframe..."
peaks_w_branch = peak_branch_analysis(peak_df, branch_bams, gff3, fa_dict, organism, base_dir, name)
else:
peaks_w_branch = pd.read_pickle(base_dir+name+'_peaks_w_branch.pickle')
print "Peaks vs. branches dataframe already exists"
#### Clean up dataframe for quantitation
if name+'_quantitation.csv' not in os.listdir(base_dir):
quant_df, lariat_df = SP.make_quant_df(peaks_w_junc, peaks_w_branch, gff3, fa_dict, organism=organism)
quant_df = SP.find_score_branches_ppy(quant_df, peaks_w_branch, fa_dict)
print "Counting reads in transcripts and at peaks..."
quant_df = SP.quantitate_junction_df(quant_bams, quant_df, gff3, organism=organism)
quant_df.to_pickle(base_dir+name+'_quantitation.pickle')
quant_df.to_csv(base_dir+name+'_quantitation.csv')
lariat_df.to_pickle(base_dir+name+'_lariats.pickle')
lariat_df.to_csv(base_dir+name+'_lariats.csv')
scatter = SP.SP_pipeline_scatters(quant_df, base_dir, name)
else:
quant_df = pd.read_pickle(base_dir+name+'_quantitation.pickle')
scatter = SP.SP_pipeline_scatters(quant_df, base_dir, name)
print "\n****Finished****"
def position_wise_scores2(seq5_list,
seq3_list,
organism,
title='Intron position strength'):
'''Uses chi-contingency test to score base proportions at each position in sample against population'''
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
all_5p, all_3p = generate_all_ss_seqs(gff3, fa_dict, organism)
pop_5p = seq_list_to_totals(all_5p)
pop_3p = seq_list_to_totals(all_3p)
samp_5p = seq_list_to_totals(seq5_list)
samp_3p = seq_list_to_totals(seq3_list)
print samp_5p.shape
p5 = []
for n in range(samp_5p.shape[1]):
if n == 2 or n == 3:
p5.append(1)
else:
conting = np.array([samp_5p[:, n], pop_5p[:, n]])
chi2, p, dof, expected = stats.chi2_contingency(conting)
p5.append(np.log10(p) * -1)
p3 = []
for n in range(samp_3p.shape[1]):
if n == 4 or n == 5:
p3.append(1)
else:
conting = np.array([samp_3p[:, n], pop_3p[:, n]])
chi2, p, dof, expected = stats.chi2_contingency(conting)
p3.append(np.log10(p) * -1)
fig, ax = plt.subplots(2, 1, figsize=(4, 4))
width = 0.7
max_y = max(p5 + p3) + 0.1 * max(p5 + p3)
ind5 = np.arange(len(p5))
ax[0].bar(ind5, p5, color='k')
ax[0].plot([0, 8], [2, 2], '--', color='0.7')
ax[0].set_xlim([0, len(p5)])
ax[0].set_ylabel("5' splice site\n-log10(p-value)")
ax[0].set_title(title)
ax[0].set_ylim([0, max_y])
ind3 = np.arange(len(p3))
ax[1].bar(ind3, p3, color='k')
ax[1].plot([0, 8], [2, 2], '--', color='0.7')
ax[1].set_xlim([0, len(p3)])
ax[1].set_ylabel("3' splice site\n-log10(p-value)")
ax[1].set_ylim([0, max_y])
ax[0].set_xticks(ind3 + width / 2)
ax[1].set_xticks(ind3 + width / 2)
ax[0].set_xticklabels(np.arange(-2, 6))
ax[1].set_xticklabels(np.arange(-5, 3))
fig.tight_layout()
plt.show()
return fig
def main():
'''Usage: run SPBranch.py unmapped1 unmapped2 threads organism [config_file] [untagged]
Parameters
-----------
unmapped1 : bam or fastq file of unmapped reads from tophat or bowtie
unmapped2 : bam or fastq file of unmapped reads from tophat or bowtie
threads : number of processors to use
organism : 'pombe or 'crypto'
config_file : if using peaks to call - list of changepoint output file names and where to find them
untagged : untagged sample name (must be in file name)
Output
------
bam files with aligned reads. Will be interpreted by SP_pipeline.
'''
unmapped1 = sys.argv[1]
unmapped2 = sys.argv[2]
threads = int(sys.argv[3])
if unmapped1.endswith('bam'):
btf_args = 'bamToFastq -i {0} -fq {1}'.format(unmapped1, unmapped1.split('.bam')[-1]+'.fq')
call(btf_args, shell=False)
unmapped1 = unmapped1.split('.bam')[-1]+'.fq'
if unmapped2.endswith('bam'):
btf_args = 'bamToFastq -i {0} -fq {1}'.format(unmapped2, unmapped2.split('.bam')[-1]+'.fq')
call(btf_args, shell=False)
unmapped2 = unmapped2.split('.bam')[-1]+'.fq'
cat_args = 'cat {0} {1} > unmapped_all.fq'.format(unmapped1, unmapped2)
call(cat_args, shell=True)
organism = sys.argv[4]
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
peaks = False
if len(sys.argv) == 7:
peaks = True
with open(sys.argv[5], 'r') as config:
for line in config:
if sys.argv[6] in line:
CP_untagged = line.strip()
elif 'changepoint' in line.lower() or 'peak' in line.lower():
CP_out.append(line.strip())
peak_df = SP.peak_to_seq_pipeline(CP_untagged, CP_out[0], CP_out[1], gff3, fa_dict, name='CP_peaks')
ann_seqs = collect_intron_seq(gff3, fa_dict)
print "Finding unaligned reads with annotated 5' splice sites"
find_split_reads('unmapped_all.fq', ann_seqs, 'Ann_branches', threads=threads)
print "Aligning split reads to the genome with Bowtie"
bowtie_args = 'bowtie -p{0} -v1 -M1 --best {1} -f Ann_branches_split.fa --sam Ann_branches.sam'.format(threads, bowtie_index)
call(bowtie_args, shell=True)
# sort and index
print "Sorting and indexing bam file"
samtools1 = 'samtools view -Sbo Ann_branches.bam Ann_branches.sam'
call(samtools1, shell=True)
samtools2 = 'samtools sort Ann_branches.bam -o Ann_branches_sorted.bam'
call(samtools2, shell=True)
samtools3 = 'samtools index Ann_branches_sorted.bam'
call(samtools3, shell=True)
if peaks is True:
print "Finding unaligned reads with unpredicted splicing events"
peak_seqs = collect_intron_seq(gff3, fa_dict, peak_df=peak_df)
find_split_reads('Ann_branches_unsplit.fa', peak_seqs, 'Peak_branches', threads=threads)
print "Aligning split reads to the genome with Bowtie"
bowtie_args = 'bowtie -p{0} -v1 -M1 --best {1} -f Peak_branches_split.fa --sam Peak_branches.sam'.format(threads, bowtie_index)
call(bowtie_args, shell=True)
print "Sorting and indexing bam file"
samtools1 = 'samtools view -Sbo Peak_branches.bam Peak_branches.sam'
call(samtools1, shell=True)
samtools2 = 'samtools sort Peak_branches.bam -o Peak_branches_sorted.bam'
call(samtools2, shell=True)
samtools3 = 'samtools index Peak_branches_sorted.bam'
call(samtools3, shell=True)
def igv_plots_general(bam_list, gene_list, organism, colors=None, names=None, save_dir=None,
unstranded=False, end_only=False, same_yaxis=False, specific_range=None, transcript_direction=True,
log_scale=False, rpm=True, PE=False, plot_junctions=False):
'''Usage:
Parameters
----------
bam_list : list, bam files in order of plotting (top to bottom)
gene_list : list of transcripts to plot (should be genes not transcript isoforms)
if dataframe passed instead of list, will plot introns (must have intron information in datafame)
organism : str, pombe or crypto
colors : list, default `None`
list of colors to use, same length as bam_list, check matplotlib documentation for valid color names
names : list, default `None`
list of sample names to use instead of bam file names. Same length as bam_files
save_dir : str, default `None`
directory to save eps files. If None, does not save files
unstranded : bool, default `False`
Use True for ChIP or DNA sequencing data (or unstranded RNAseq)
end_only : bool or list, default `False`
Whether to plot only the ends of reads. If different for each bam, make a list of bools same length as bam_list
same_yaxis : bool, default `False`
Whether all samples should be plotted on the same axis after normalizing to total number of aligned reads
specific_range : str, default `None`
Options: ('end', window)
('start', window)
([coordinate], window)
transcript_direction : bool, default `True`
If True, will plot in the direction of transcription, not in the direction of the DNA
'''
# Get all organism information (annotation etc.)
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
tx_dict = SP.build_transcript_dict(gff3, organism=organism)
fix_info = {'I':'chr1','II':'chr2','III':'chr3','chr1':'I','chr2':'II','chr4':'IV','chr5':'V','chr6':'VI',
'chr7':'VII','chr8':'VIII','chr9':'IX','chr10':'X','chr11':'XI','chr12':'XII','chr13':'XIII',
'chr14':'XIV','chr15':'XV','chr16':'XVI','-':'+','+':'-','chr1':'I','chr2':'II','chr3':'III'}
if organism == 'pombe':
tx_suffix = '.1'
else:
tx_suffix = 'T0'
# Set up range parameters if specific range is indicated
if specific_range is not None:
window = int(specific_range[1])
new_tx_dict = {}
for gene in gene_list:
info = tx_dict[gene+tx_suffix]
if specific_range[0] == 'end':
if info[2] == '+':
start = info[1]-window
end = info[1]+window
else:
start = info[0]-window
end = info[0]+window
elif specific_range[0] == 'start':
if info[2] == '-':
start = info[1]-window
end = info[1]+window
else:
start = info[0]-window
end = info[0]+window
else:
start = int(specific_range[0])-window
end = int(specific_range[0])+window
new_tx_dict[gene+tx_suffix] = [start, end, info[2], info[3]]
else:
new_tx_dict = tx_dict
# Open bam files and count reads if rpm is True
open_bams = {}
total_list = []
for bam in bam_list:
open_bams[bam] = pysam.Samfile(bam)
if rpm is True:
total = check_output(['samtools','view','-F 0x04','-c',bam]).strip()
total = float(total)/1000000.
total_list.append(total)
else:
total_list.append(1.)
# Expand optional arguments to lists if necessary
colors = list_from_arg(colors, len(bam_list))
end_only = list_from_arg(end_only, len(bam_list))
log_scale = list_from_arg(log_scale, len(bam_list))
unstranded = list_from_arg(unstranded, len(bam_list))
# Get gene_list from dataframe if gene_list is not a list
df = None
if type(gene_list) == dict:
new_tx_dict = gene_list
gene_list = gene_list.keys()
elif type(gene_list) != list:
df = gene_list
gene_list = df.index
for tx in gene_list:
num_ax = len(bam_list)+1
if plot_junctions is True:
num_ax += len(bam_list)
fig, ax = plt.subplots(num_ax, figsize=(10,num_ax), sharex=True)
fig.subplots_adjust(hspace=0)
# Get transcript info from transcript_dictionary
if df is None:
try:
info = new_tx_dict[tx+tx_suffix]
except KeyError:
info = new_tx_dict[tx]
chrom = info[3]
start = info[0]
end = info[1]
strand = info[2]
# If dataframe was passed, get plotting information from dataframe instead
else:
if isinstance(df.columns, pd.core.index.MultiIndex):
new_columns = [x[1] for x in df.columns if x[0] == 'Peaks']
df = df[[x for x in df.columns if x[0] == 'Peaks']]
df.columns = new_columns
strand = df.loc[tx,'strand']
chrom = df.loc[tx,'chromosome']
if strand == '+':
start = df.loc[tx,'position']-100
end = df.loc[tx,'position'] + df.loc[tx,'intron size']+100
elif strand == '-':
start = df.loc[tx,'position']-df.loc[tx,'intron size']-100
end = df.loc[tx,'position']+100
start = int(start)
end = int(end)
tx = df.loc[tx,'transcript']
# Generate read series for each transcript
max_y = 0
junc_ymax = 0
for n, bam in enumerate(bam_list):
try:
bam_iter = open_bams[bam].fetch(chrom, start, end)
except ValueError:
chrom = fix_info[chrom]
bam_iter = open_bams[bam].fetch(chrom, start, end)
if end_only[n] is not False:
s = SP.generate_read_series_A(bam_iter, chrom, start, end, strand)
linewidth = 2
else:
if PE is False:
s = SP.generate_read_series_B(bam_iter, chrom, start, end, strand)
else:
s = SP.generate_read_series_PE(bam_iter, chrom, start, end, strand)
linewidth = 1
# Get reads from otherstrand if the library type is unstranded
if unstranded[n] is True:
bam_iter = open_bams[bam].fetch(chrom, start, end)
if end_only[n] is not False:
s2 = SP.generate_read_series_A(bam_iter, chrom, start, end, fix_info[strand])
linewidth = 2
else:
if PE is False:
s2 = SP.generate_read_series_B(bam_iter, chrom, start, end, fix_info[strand])
else:
s2 = SP.generate_read_series_PE(bam_iter, chrom, start, end, fix_info[strand])
linewidth = 1
s = s.add(s2)
# Normalize to rpm (will just divide by 1 if rpm is False)
s = s.divide(total_list[n])
if log_scale[n] is True:
s = s.apply(np.log2)
# Plot!
ax[n].bar(s.index, s, linewidth=linewidth, color=colors[n], edgecolor=colors[n], zorder=2)
ax[n].tick_params(axis='both', which='major', labelsize=14)
max_y = max([max_y,max(s)])
if plot_junctions is True:
m = n+len(bam_list)
intron_dict = get_junctions(open_bams[bam], chrom, start, end, strand)
ax[m].plot((start, end),(0,0),'-',c='k')
for coords, heights in intron_dict.iteritems():
ax[m].plot(coords, heights, '-', linewidth=2, color=colors[n])
ax[m].fill_between(coords, 0, heights, facecolor=colors[n], interpolate=True, alpha=0.5)
if same_yaxis is True:
junc_ymax = max([junc_ymax, max(zip(*intron_dict.values())[1])])
# Add diagram of gene below traces
if tx in tx_dict:
strand = gene_patches(tx, tx_dict, ax[-1])
ax[-1].set_xlim(start, end)
else:
try:
new_tx = tx.split(' ')[0]
if new_tx[-2] == 'T' or new_tx[-2] == '.':
new_tx = new_tx[:-2]
strand = gene_patches(new_tx, tx_dict, ax[-1])
ax[-1].set_xlim(start, end)
except KeyError:
print "Transcript unknown"
# Flip minus strand transcripts if indicated
if transcript_direction is True:
if strand == '-':
ax[-1].invert_xaxis()
# Set x and y limits
for n in range(len(bam_list)):
ax[n].set_xlim(start, end)
if same_yaxis is True:
ax[n].set_ylim(0,max_y+0.1*max_y)
if plot_junctions is True:
ax[n+len(bam_list)].set_ylim(0,junc_ymax+0.1*junc_ymax)
if strand == '-':
ax[n].invert_xaxis()
ax[0].set_ylabel('RPM', fontsize=16)
ax[0].set_title(tx, fontsize=16)
#ax[0].get_xaxis().set_ticks([])
plt.show()
# Save if indicated
if save_dir is not None:
if not os.path.exists(save_dir):
os.makedirs(save_dir)
fig.savefig(save_dir+tx+'.eps', format='eps')
plt.clf()
def main():
'''Usage: run SPBranch.py unmapped1 unmapped2 threads organism [config_file] [untagged]
Parameters
-----------
unmapped1 : bam or fastq file of unmapped reads from tophat or bowtie
unmapped2 : bam or fastq file of unmapped reads from tophat or bowtie
threads : number of processors to use
organism : 'pombe or 'crypto'
config_file : if using peaks to call - list of changepoint output file names and where to find them
untagged : untagged sample name (must be in file name)
Output
------
bam files with aligned reads. Will be interpreted by SP_pipeline.
'''
unmapped1 = sys.argv[1]
unmapped2 = sys.argv[2]
threads = int(sys.argv[3])
if unmapped1.endswith('bam'):
btf_args = 'bamToFastq -i {0} -fq {1}'.format(
unmapped1,
unmapped1.split('.bam')[-1] + '.fq')
call(btf_args, shell=False)
unmapped1 = unmapped1.split('.bam')[-1] + '.fq'
if unmapped2.endswith('bam'):
btf_args = 'bamToFastq -i {0} -fq {1}'.format(
unmapped2,
unmapped2.split('.bam')[-1] + '.fq')
call(btf_args, shell=False)
unmapped2 = unmapped2.split('.bam')[-1] + '.fq'
cat_args = 'cat {0} {1} > unmapped_all.fq'.format(unmapped1, unmapped2)
call(cat_args, shell=True)
organism = sys.argv[4]
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
peaks = False
if len(sys.argv) == 7:
peaks = True
with open(sys.argv[5], 'r') as config:
for line in config:
if sys.argv[6] in line:
CP_untagged = line.strip()
elif 'changepoint' in line.lower() or 'peak' in line.lower():
CP_out.append(line.strip())
peak_df = SP.peak_to_seq_pipeline(CP_untagged,
CP_out[0],
CP_out[1],
gff3,
fa_dict,
name='CP_peaks')
ann_seqs = collect_intron_seq(gff3, fa_dict)
print "Finding unaligned reads with annotated 5' splice sites"
find_split_reads('unmapped_all.fq',
ann_seqs,
'Ann_branches',
threads=threads)
print "Aligning split reads to the genome with Bowtie"
bowtie_args = 'bowtie -p{0} -v1 -M1 --best {1} -f Ann_branches_split.fa --sam Ann_branches.sam'.format(
threads, bowtie_index)
call(bowtie_args, shell=True)
# sort and index
print "Sorting and indexing bam file"
samtools1 = 'samtools view -Sbo Ann_branches.bam Ann_branches.sam'
call(samtools1, shell=True)
samtools2 = 'samtools sort Ann_branches.bam -o Ann_branches_sorted.bam'
call(samtools2, shell=True)
samtools3 = 'samtools index Ann_branches_sorted.bam'
call(samtools3, shell=True)
if peaks is True:
print "Finding unaligned reads with unpredicted splicing events"
peak_seqs = collect_intron_seq(gff3, fa_dict, peak_df=peak_df)
find_split_reads('Ann_branches_unsplit.fa',
peak_seqs,
'Peak_branches',
threads=threads)
print "Aligning split reads to the genome with Bowtie"
bowtie_args = 'bowtie -p{0} -v1 -M1 --best {1} -f Peak_branches_split.fa --sam Peak_branches.sam'.format(
threads, bowtie_index)
call(bowtie_args, shell=True)
print "Sorting and indexing bam file"
samtools1 = 'samtools view -Sbo Peak_branches.bam Peak_branches.sam'
call(samtools1, shell=True)
samtools2 = 'samtools sort Peak_branches.bam -o Peak_branches_sorted.bam'
call(samtools2, shell=True)
samtools3 = 'samtools index Peak_branches_sorted.bam'
call(samtools3, shell=True)
def peak_seq_enrichment(df, organism):
organism, gff3, fa_dict, bowtie_index = SP.find_organism_files(organism)
nuc_prob = SP.gc_content(fa_dict)
p_dict = {
'A': nuc_prob[0],
'T': nuc_prob[2],
'C': nuc_prob[1],
'G': nuc_prob[3]
}
unpeaks = df[df['type'] == 'other']
unpeaks = unpeaks.append(df[df['type'] == 'intronic'])
print "Number of unpredicted peaks:"
print len(unpeaks)
nucs = ['G', 'A', 'C', 'T']
dinucs = set()
for nuc in nucs:
for nuc2 in nucs:
dinucs.add(nuc + nuc2)
five = {}
three = {}
for dinuc in dinucs:
five[dinuc] = len(
unpeaks[unpeaks['sequence'].str[6:8].str.contains(dinuc)])
three[dinuc] = len(
unpeaks[unpeaks['sequence'].str[4:6].str.contains(dinuc)])
five_LO = {}
three_LO = {}
for dinuc in five.keys():
p_dinuc = p_dict[dinuc[0]] * p_dict[dinuc[1]]
phat_dinuc = five[dinuc] / float(len(unpeaks))
phat_dinuc2 = three[dinuc] / float(len(unpeaks))
SE = np.sqrt(phat_dinuc * (1 - phat_dinuc) / len(unpeaks))
SE2 = np.sqrt(phat_dinuc2 * (1 - phat_dinuc2) / len(unpeaks))
Z = (phat_dinuc - p_dinuc) / SE
Z2 = (phat_dinuc2 - p_dinuc) / SE2
pvalue = stats.norm.sf(Z)
pvalue2 = stats.norm.sf(Z2)
LO = np.log((1 - pvalue) / pvalue)
LO2 = np.log((1 - pvalue2) / pvalue2)
five_LO[dinuc] = LO
three_LO[dinuc] = LO2
fig, ax = plt.subplots(figsize=(12, 6))
width = 0.35
ind = np.arange(len(five_LO.keys()))
rects2 = ax.bar(ind,
three_LO.values(),
width,
color='crimson',
edgecolor='crimson',
label='Before peak')
rects1 = ax.bar(ind + width,
five_LO.values(),
width,
color='indigo',
edgecolor='indigo',
label='After peak')
ax.plot([-1, 17], [0, 0], '-', color='black')
ax.plot([-1, 17], [2.94, 2.94], '--', color='0.7', label='95% CI')
ax.plot([-1, 17], [-2.94, -2.94], '--', color='0.7')
ax.set_xlim([-1, 17])
ax.set_xticklabels(five_LO.keys(), fontsize=12)
ax.set_xticks(ind + width / 2)
ax.set_ylabel('Log odds dinucleotide enrichment', fontsize=14)
ax.set_title('Unpredicted peaks', fontsize=14)
ax.legend(fontsize=12)
return fig
