def wrapper(args):
try:
npar = args.noiseparam.strip("[").strip("]").split(",")
except:
npar = []
nbins = args.nbins
# Run funciton
if args.i:
df = pd.io.parsers.read_csv(args.i, delim_whitespace=True, dtype={"seqs": str, "batch": int})
else:
df = pd.io.parsers.read_csv(sys.stdin, delim_whitespace=True, dtype={"seqs": str, "batch": int})
if len(utils.get_column_headers(df)) > 0:
raise SortSeqError("Library already sorted!")
model_df = io.load_model(args.model)
output_df = main(df, model_df, args.noisemodel, npar, nbins, start=args.start, end=args.end)
if args.out:
outloc = open(args.out, "w")
else:
outloc = sys.stdout
pd.set_option("max_colwidth", int(1e8))
# Validate dataframe for writting
output_df = qc.validate_dataset(output_df, fix=True)
io.write(output_df, outloc)
def wrapper(args):
T_LibCounts = args.totallibcounts
T_mRNACounts = args.totalmRNAcounts
if T_LibCounts <=0 or T_mRNACounts <= 0:
raise SortSeqError('Counts must be greater than zero')
model_df = io.load_model(args.model)
if args.i:
df = pd.io.parsers.read_csv(args.i,delim_whitespace=True)
else:
df = pd.io.parsers.read_csv(sys.stdin,delim_whitespace=True)
#make sure the library is not already sorted
if len(utils.get_column_headers(df)) > 0:
raise SortSeqError('Library already sorted!')
header = df.columns
libcounts,expcounts = main(df,model_df,T_LibCounts,T_mRNACounts,start=args.start,end=args.end)
#add these counts to input dataframe
lc = pd.Series(libcounts,name='ct_0')
ec = pd.Series(expcounts,name='ct_1')
df['ct_0'] = lc
df['ct_1'] = ec
df['ct'] = df[['ct_0','ct_1']].sum(axis=1)
if args.out:
outloc = open(args.out,'w')
else:
outloc = sys.stdout
pd.set_option('max_colwidth',int(1e8))
# Validate dataframe for writting
df = qc.validate_dataset(df,fix=True)
io.write(df,outloc)
def main(
data_df,model_df,
start=0,end=None,err=False,coarse_graining_level=0):
dicttype, modeltype = qc.get_model_type(model_df)
seq_cols = qc.get_cols_from_df(data_df,'seqs')
if not len(seq_cols)==1:
raise SortSeqError('Dataframe has multiple seq cols: %s'%str(seq_cols))
seq_dict,inv_dict = utils.choose_dict(dicttype,modeltype=modeltype)
#set name of sequences column based on type of sequence
type_name_dict = {'dna':'seq','rna':'seq_rna','protein':'seq_pro'}
seq_col_name = type_name_dict[dicttype]
#Cut the sequences based on start and end, and then check if it makes sense
if (start != 0 or end):
data_df.loc[:,seq_col_name] = data_df.loc[:,seq_col_name].str.slice(start,end)
if modeltype=='MAT':
if len(data_df.loc[0,seq_col_name]) != len(model_df.loc[:,'pos']):
raise SortSeqError('model length does not match dataset length')
elif modeltype=='NBR':
if len(data_df.loc[0,seq_col_name]) != len(model_df.loc[:,'pos'])+1:
raise SortSeqError('model length does not match dataset length')
col_headers = utils.get_column_headers(data_df)
if 'ct' not in data_df.columns:
data_df['ct'] = data_df[col_headers].sum(axis=1)
data_df = data_df[data_df.ct != 0]
if not end:
seqL = len(data_df[seq_col_name][0]) - start
else:
seqL = end-start
data_df = data_df[data_df[seq_col_name].apply(len) == (seqL)]
#make a numpy array out of the model data frame
model_df_headers = ['val_' + str(inv_dict[i]) for i in range(len(seq_dict))]
value = np.transpose(np.array(model_df[model_df_headers]))
#now we evaluate the expression of each sequence according to the model.
seq_mat,wtrow = numerics.dataset2mutarray(data_df.copy(),modeltype)
temp_df = data_df.copy()
temp_df['val'] = numerics.eval_modelmatrix_on_mutarray(np.array(model_df[model_df_headers]),seq_mat,wtrow)
temp_sorted = temp_df.sort_values(by='val')
temp_sorted.reset_index(inplace=True,drop=True)
#we must divide by the total number of counts in each bin for the MI calculator
#temp_sorted[col_headers] = temp_sorted[col_headers].div(temp_sorted['ct'],axis=0)
MI = EstimateMutualInfoforMImax.alt4(temp_sorted,coarse_graining_level=coarse_graining_level)
if not err:
Std = np.NaN
else:
data_df_for_sub = data_df.copy()
sub_MI = np.zeros(15)
for i in range(15):
sub_df = data_df_for_sub.sample(int(len(data_df_for_sub.index)/2))
sub_df.reset_index(inplace=True,drop=True)
sub_MI[i],sub_std = main(
sub_df,model_df,err=False)
Std = np.std(sub_MI)/np.sqrt(2)
return MI,Std
def Berg_von_Hippel(df,dicttype,foreground=1,background=0,pseudocounts=1):
'''Learn models using berg von hippel model. The foreground sequences are
usually bin_1 and background in bin_0, this can be changed via flags.'''
seq_dict,inv_dict = utils.choose_dict(dicttype)
#check that the foreground and background chosen columns actually exist.
columns_to_check = {'ct_' + str(foreground),'ct_' + str(background)}
if not columns_to_check.issubset(set(df.columns)):
raise SortSeqError('Foreground or Background column does not exist!')
#get counts of each base at each position
foreground_counts = utils.profile_counts(df,dicttype,bin_k=foreground)
background_counts = utils.profile_counts(df,dicttype,bin_k=background)
binheaders = utils.get_column_headers(foreground_counts)
#add pseudocounts to each position
foreground_counts[binheaders] = foreground_counts[binheaders] + pseudocounts
background_counts[binheaders] = background_counts[binheaders] + pseudocounts
#make sure there are no zeros in counts after addition of pseudocounts
ct_headers = utils.get_column_headers(foreground_counts)
if foreground_counts[ct_headers].isin([0]).values.any():
raise SortSeqError('''There are some bases without any representation in\
the foreground data, you should use pseudocounts to avoid failure \
of the learning method''')
if background_counts[ct_headers].isin([0]).values.any():
raise SortSeqError('''There are some bases without any representation in\
the background data, you should use pseudocounts to avoid failure \
of the learning method''')
#normalize to compute frequencies
foreground_freqs = foreground_counts.copy()
background_freqs = background_counts.copy()
foreground_freqs[binheaders] = foreground_freqs[binheaders].div(
foreground_freqs[binheaders].sum(axis=1),axis=0)
background_freqs[binheaders] = background_freqs[binheaders].div(
background_freqs[binheaders].sum(axis=1),axis=0)
output_df = -np.log(foreground_freqs/background_freqs)
#change column names accordingly (instead of ct_ we want val_)
rename_dict = {'ct_' + str(inv_dict[i]):'val_' + str(inv_dict[i]) for i in range(len(seq_dict))}
output_df = output_df.rename(columns=rename_dict)
return output_df
def main(df, mp, noisetype, npar, nbins, sequence_library=True, start=0, end=None):
# validate noise parameters
if not isinstance(npar, list):
raise SortSeqError("Noise parameters must be given as a list")
if noisetype == "Normal":
if len(npar) != 1:
raise SortSeqError(
"""For a normal noise model, there must be one
input parameter (width of normal distribution)"""
)
if noisetype == "LogNormal":
if len(npar) != 2:
raise SortSeqError(
"""For a LogNormal noise model there must
be 2 input parameters"""
)
if nbins <= 1:
raise SortSeqError("number of bins must be greater than 1")
# generate predicted energy of each sequence.
df = evaluate_model.main(df, mp, left=start, right=None)
# Determine model type to use for noise
if noisetype == "LogNormal":
NoiseModelSort = Models.LogNormalNoise(npar)
elif noisetype == "Normal":
NoiseModelSort = Models.NormalNoise(npar)
elif noisetype == "None":
NoiseModelSort = Models.NormalNoise([1e-16])
else:
NoiseModelSort = Models.CustomModel(noisetype, npar)
# Apply noise to our calculated energies
noisyexp, listnoisyexp = NoiseModelSort.genlist(df)
# Determine Expression Cutoffs for bins
noisyexp.sort()
cutoffs = list(noisyexp[np.linspace(0, len(noisyexp), nbins, endpoint=False, dtype=int)])
cutoffs.append(np.inf)
seqs_arr = np.zeros([len(listnoisyexp), nbins], dtype=int)
# split sequence into bins based on calculated cutoffs
for i, entry in enumerate(listnoisyexp):
seqs_arr[i, :] = np.histogram(entry, bins=cutoffs)[0]
col_labels = ["ct_" + str(i + 1) for i in range(nbins)]
if sequence_library:
df["ct_0"] = utils.sample(df["ct"], int(df["ct"].sum() / nbins))
output_df = pd.concat([df, pd.DataFrame(seqs_arr, columns=col_labels)], axis=1)
col_labels = utils.get_column_headers(output_df)
output_df["ct"] = output_df[col_labels].sum(axis=1)
output_df = output_df.drop("val", axis=1)
return output_df
def main(df,lm='IM',modeltype='MAT',LS_means_std=None,\
db=None,iteration=30000,burnin=1000,thin=10,\
runnum=0,initialize='LS',start=0,end=None,foreground=1,\
background=0,alpha=0,pseudocounts=1,test=False,drop_library=False,\
verbose=False):
# Determine dictionary
seq_cols = qc.get_cols_from_df(df,'seqs')
if not len(seq_cols)==1:
raise SortSeqError('Dataframe has multiple seq cols: %s'%str(seq_cols))
dicttype = qc.colname_to_seqtype_dict[seq_cols[0]]
seq_dict,inv_dict = utils.choose_dict(dicttype,modeltype=modeltype)
'''Check to make sure the chosen dictionary type correctly describes
the sequences. An issue with this test is that if you have DNA sequence
but choose a protein dictionary, you will still pass this test bc A,C,
G,T are also valid amino acids'''
#set name of sequences column based on type of sequence
type_name_dict = {'dna':'seq','rna':'seq_rna','protein':'seq_pro'}
seq_col_name = type_name_dict[dicttype]
lin_seq_dict,lin_inv_dict = utils.choose_dict(dicttype,modeltype='MAT')
#wtseq = utils.profile_counts(df.copy(),dicttype,return_wtseq=True,start=start,end=end)
#wt_seq_dict_list = [{inv_dict[np.mod(i+1+seq_dict[w],len(seq_dict))]:i for i in range(len(seq_dict)-1)} for w in wtseq]
par_seq_dict = {v:k for v,k in seq_dict.items() if k != (len(seq_dict)-1)}
#drop any rows with ct = 0
df = df[df.loc[:,'ct'] != 0]
df.reset_index(drop=True,inplace=True)
#If there are sequences of different lengths, then print error but continue
if len(set(df[seq_col_name].apply(len))) > 1:
sys.stderr.write('Lengths of all sequences are not the same!')
#select target sequence region
df.loc[:,seq_col_name] = df.loc[:,seq_col_name].str.slice(start,end)
df = utils.collapse_further(df)
col_headers = utils.get_column_headers(df)
#make sure all counts are ints
df[col_headers] = df[col_headers].astype(int)
#create vector of column names
val_cols = ['val_' + inv_dict[i] for i in range(len(seq_dict))]
df.reset_index(inplace=True,drop=True)
#Drop any sequences with incorrect length
if not end:
'''is no value for end of sequence was supplied, assume first seq is
correct length'''
seqL = len(df[seq_col_name][0]) - start
else:
seqL = end-start
df = df[df[seq_col_name].apply(len) == (seqL)]
df.reset_index(inplace=True,drop=True)
#Do something different for each type of learning method (lm)
if lm == 'ER':
emat = Berg_von_Hippel(
df,dicttype,foreground=foreground,background=background,
pseudocounts=pseudocounts)
if lm == 'LS':
'''First check that is we don't have a penalty for ridge regression,
that we at least have all possible base values so that the analysis
will not fail'''
if LS_means_std: #If user supplied preset means and std for each bin
means_std_df = io.load_meanstd(LS_means_std)
#change bin number to 'ct_number' and then use as index
labels = list(means_std_df['bin'].apply(add_label))
std = means_std_df['std']
std.index = labels
#Change Weighting of each sequence by dividing counts by bin std
df[labels] = df[labels].div(std)
means = means_std_df['mean']
means.index = labels
else:
means = None
#drop all rows without counts
df['ct'] = df[col_headers].sum(axis=1)
df = df[df.ct != 0]
df.reset_index(inplace=True,drop=True)
''' For sort-seq experiments, bin_0 is library only and isn't the lowest
expression even though it is will be calculated as such if we proceed.
Therefore is drop_library is passed, drop this column from analysis.'''
if drop_library:
try:
df.drop('ct_0',inplace=True)
col_headers = utils.get_column_headers(df)
if len(col_headers) < 2:
raise SortSeqError(
'''After dropping library there are no longer enough
columns to run the analysis''')
except:
raise SortSeqError('''drop_library option was passed, but no ct_0
column exists''')
#parameterize sequences into 3xL vectors
raveledmat,batch,sw = utils.genweightandmat(
df,par_seq_dict,dicttype,means=means,modeltype=modeltype)
#Use ridge regression to find matrix.
emat = Compute_Least_Squares(raveledmat,batch,sw,alpha=alpha)
if lm == 'IM':
seq_mat,wtrow = numerics.dataset2mutarray(df.copy(),modeltype)
#this is also an MCMC routine, do the same as above.
if initialize == 'rand':
if modeltype == 'MAT':
emat_0 = utils.RandEmat(len(df[seq_col_name][0]),len(seq_dict))
elif modeltype == 'NBR':
emat_0 = utils.RandEmat(len(df['seq'][0])-1,len(seq_dict))
elif initialize == 'LS':
emat_cols = ['val_' + inv_dict[i] for i in range(len(seq_dict))]
emat_0_df = main(df.copy(),lm='LS',modeltype=modeltype,alpha=alpha,start=0,end=None,verbose=verbose)
emat_0 = np.transpose(np.array(emat_0_df[emat_cols]))
#pymc doesn't take sparse mat
emat = MaximizeMI_memsaver(
seq_mat,df.copy(),emat_0,wtrow,db=db,iteration=iteration,burnin=burnin,
thin=thin,runnum=runnum,verbose=verbose)
#now format the energy matrices to get them ready to output
if (lm == 'IM' or lm == 'memsaver'):
if modeltype == 'NBR':
emat_typical = gauge.fix_neighbor(np.transpose(emat))
elif modeltype == 'MAT':
emat_typical = gauge.fix_matrix(np.transpose(emat))
elif lm == 'ER':
'''the emat for this format is currently transposed compared to other formats
it is also already a data frame with columns [pos,val_...]'''
emat_cols = ['val_' + inv_dict[i] for i in range(len(seq_dict))]
emat_typical = emat[emat_cols]
emat_typical = (gauge.fix_matrix((np.array(emat_typical))))
else: #must be Least squares
emat_typical = utils.emat_typical_parameterization(emat,len(seq_dict))
if modeltype == 'NBR':
emat_typical = gauge.fix_neighbor(np.transpose(emat_typical))
elif modeltype == 'MAT':
emat_typical = gauge.fix_matrix(np.transpose(emat_typical))
em = pd.DataFrame(emat_typical)
em.columns = val_cols
#add position column
if modeltype == 'NBR':
pos = pd.Series(range(start,start - 1 + len(df[seq_col_name][0])),name='pos')
else:
pos = pd.Series(range(start,start + len(df[seq_col_name][0])),name='pos')
output_df = pd.concat([pos,em],axis=1)
# Validate model and return
output_df = qc.validate_model(output_df,fix=True)
return output_df
def main(data_df,
model_df,
start=0,
end=None,
err=False,
coarse_graining_level=0,
rsquared=False,
return_freg=False):
#determine whether you are working with RNA, DNA, or protein.
#this also should determine modeltype (MAT, NBR, PAIR).
dicttype, modeltype = qc.get_model_type(model_df)
#get column header for the sequence column.
seq_cols = qc.get_cols_from_df(data_df, 'seqs')
if not len(seq_cols) == 1:
raise SortSeqError('Dataframe has multiple seq cols: %s' %
str(seq_cols))
#create dictionary that goes from, for example, nucleotide to number and
#visa versa.
seq_dict, inv_dict = utils.choose_dict(dicttype, modeltype=modeltype)
#set name of sequences column based on type of sequence
type_name_dict = {'dna': 'seq', 'rna': 'seq_rna', 'protein': 'seq_pro'}
seq_col_name = type_name_dict[dicttype]
if not end:
seqL = len(data_df[seq_col_name][0]) - start
else:
seqL = end - start
#throw out wrong length sequences.
#Cut the sequences based on start and end, and then check if it makes sense
if (start != 0 or end):
data_df.loc[:,seq_col_name] = \
data_df.loc[:,seq_col_name].str.slice(start,end)
right_length = data_df.loc[:, seq_col_name].apply(len) == (seqL)
if not right_length.all():
sys.stderr.write('''Not all sequences are the same length!
Throwing out incorrect sequences!''')
data_df = data_df.loc[right_length, :]
data_df = data_df.reset_index(drop=True)
if modeltype == 'MAT':
if seqL != len(model_df.loc[:, 'pos']):
raise SortSeqError(
'model length does not match dataset length')
elif modeltype == 'NBR':
if seqL != len(model_df.loc[:, 'pos']) + 1:
raise SortSeqError(
'model length does not match dataset length')
elif modeltype == 'PAIR':
if int(scipy.misc.comb(seqL, 2)) != len(model_df.loc[:, 'pos']):
raise SortSeqError(
'model length does not match dataset length')
#get column names of the counts columns (excluding total counts 'ct')
col_headers = utils.get_column_headers(data_df)
if 'ct' not in data_df.columns:
data_df['ct'] = data_df[col_headers].sum(axis=1)
#remove empty rows.
data_df = data_df[data_df.ct != 0]
#determine sequence length.
#make a numpy array out of the model data frame
model_df_headers = [
'val_' + str(inv_dict[i]) for i in range(len(seq_dict))
]
value = np.array(model_df[model_df_headers])
#now we evaluate the expression of each sequence according to the model.
#first convert to matrix representation of sequences
seq_mat, wtrow = numerics.dataset2mutarray(data_df.copy(), modeltype)
temp_df = data_df.copy()
#evaluate energy of each sequence
temp_df['val'] = numerics.eval_modelmatrix_on_mutarray(
value, seq_mat, wtrow)
#sort based on value
temp_sorted = temp_df.sort_values(by='val')
temp_sorted.reset_index(inplace=True, drop=True)
#freg is a regularized plot which show how sequences are distributed
#in energy space.
if return_freg:
fig, ax = plt.subplots()
MI, freg = EstimateMutualInfoforMImax.alt4(
temp_sorted,
coarse_graining_level=coarse_graining_level,
return_freg=return_freg)
plt.imshow(freg, interpolation='nearest', aspect='auto')
plt.savefig(return_freg)
else:
MI = EstimateMutualInfoforMImax.alt4(
temp_sorted,
coarse_graining_level=coarse_graining_level,
return_freg=return_freg)
#if we want to calculate error then use bootstrapping.
if not err:
Std = np.NaN
else:
data_df_for_sub = data_df.copy()
sub_MI = np.zeros(15)
for i in range(15):
sub_df = data_df_for_sub.sample(int(
len(data_df_for_sub.index) / 2))
sub_df.reset_index(inplace=True, drop=True)
sub_MI[i], sub_std = main(sub_df, model_df, err=False)
Std = np.std(sub_MI) / np.sqrt(2)
#we can return linfoot corrolation (rsquared) or return MI.
if rsquared:
return (1 - 2**(-2 * MI)), (1 - 2**(-2 * Std))
else:
return MI, Std
def main(dataset_df,
bin=None,
start=0,
end=None,
bins_df=None,
pseudocounts=1,
return_profile=False):
"""
Computes character frequencies (0.0 to 1.0) at each position
Arguments:
dataset_df (pd.DataFrame): A dataframe containing a valid dataset.
bin (int): A bin number specifying which counts to use
start (int): An integer specifying the sequence start position
end (int): An integer specifying the sequence end position
Returns:
freq_df (pd.DataFrame): A dataframe containing counts for each
nucleotide/amino acid character at each position.
"""
seq_cols = qc.get_cols_from_df(dataset_df, 'seqs')
if not len(seq_cols) == 1:
raise SortSeqError('Dataframe has multiple seq cols: %s' %
str(seq_cols))
dicttype = qc.colname_to_seqtype_dict[seq_cols[0]]
seq_dict, inv_dict = utils.choose_dict(dicttype)
# Validate dataset_df
qc.validate_dataset(dataset_df)
#for each bin we need to find character frequency profile, then sum over all
#bins to get activity.
#first make sure we have activities of each bin:
if not bins_df:
bins = utils.get_column_headers(dataset_df)
#in this case no activity was specified so just assume the activity
#equals bin number
activity = [float(b.split('_')[-1]) for b in bins]
else:
bins = list(bins_df['bins'])
activity = list(bins_df['activity'])
#initialize dataframe for total counts in all bins
output_ct_df = pd.DataFrame()
#initialize dataframe for running activity calculation
output_activity_df = pd.DataFrame()
for i, b in enumerate(bins):
bin_num = int(b.split('_')[-1])
# Compute counts
counts_df = profile_ct.main(dataset_df,
bin=bin_num,
start=start,
end=end)
# Create columns for profile_freqs table
ct_cols = utils.get_column_headers(counts_df)
#add_pseudocounts
counts_df[ct_cols] = counts_df[ct_cols] + pseudocounts
#add to all previous bin counts
#print output_activity_df
if i == 0:
output_ct_df = counts_df[ct_cols]
output_activity_df = counts_df[ct_cols] * activity[i]
else:
output_ct_df = output_ct_df + counts_df[ct_cols]
output_activity_df = output_activity_df + counts_df[
ct_cols] * activity[i]
#now normalize by each character at each position, this is the activity
#profile
output_activity_df = output_activity_df[ct_cols].div(output_ct_df[ct_cols])
mut_rate = profile_mut.main(dataset_df, bin=bin)
freq = profile_freq.main(dataset_df, bin=bin)
freq_cols = [x for x in freq.columns if 'freq_' in x]
#now normalize by the wt activity
wtseq = ''.join(mut_rate['wt'])
wtarr = utils.seq2mat(wtseq, seq_dict)
wt_activity = np.transpose(wtarr) * (np.array(output_activity_df[ct_cols]))
#sum this to get total
wt_activity2 = wt_activity.sum(axis=1)
delta_activity = output_activity_df.subtract(pd.Series(wt_activity2),
axis=0)
if return_profile:
#first find mutation rate according to formula in SI text
profile_delta_activity = mut_rate['mut']*np.sum(
(1-np.transpose(wtarr))*np.array(\
freq[freq_cols])*np.array(delta_activity),axis=1)
#format into dataframe
output_df = pd.DataFrame()
output_df['pos'] = range(start,
start + len(profile_delta_activity.index))
output_df['mut_activity'] = profile_delta_activity
return output_df
else:
#just add pos column and rename counts columns to activity columns
output_df = pd.DataFrame(delta_activity)
output_df.insert(0, 'pos',
range(start, start + len(delta_activity.index)))
#reorder columns
activity_col_dict = {x:'activity_' + x.split('_')[-1] \
for x in delta_activity.columns if 'ct_' in x}
output_df = output_df.rename(columns=activity_col_dict)
return output_df
