def pairs_notfound_sps(df, fs, sps="Hs Mm Sp Dm Ce".split()):
"""
df: dataframe with id1, id2, and the sp_evidence columns.
fs: all the elution filenames
"""
results = []
for sp in sps:
pairs = [(r["id1"], r["id2"]) for i, r in df[df[sp + "_evidence"] != "frac"].iterrows()]
print "%s pairs for %s" % (len(pairs), sp)
odict = orth.odict("Hs", sp)
orths = pairs_found(pairs, odict) if odict else len(pairs) # same sp
fs_sp = [f for f in fs if f.find(sp + "_") > -1]
print "%s fractionations for %s" % (len(fs_sp), sp)
allps = el.all_prots(fs_sp)
counts = pairs_orth_found(pairs, odict, allps)
results.append((len(pairs), orths, counts))
return sps, results
def prot_counts(fs, min_count=2):
"""
Sum up all the spectral counts for all the proteins in a set of
fractionations.
Filtered s.t. any returned protein will have at least min_count counts in
one fraction of one of the fractionations.
Return a dict: {prot1:count1, prot2:count2, ...}
"""
allprots = el.all_prots(fs, min_count=min_count)
pcounts = collections.defaultdict(float)
for f in fs:
e = el.load_elution(f)
psums = np.sum(np.array(e.mat),axis=1)
frac_sum = sum(psums)
norm_term = 1 / (frac_sum * len(fs))
for p,psum in zip(e.prots,psums):
if p in allprots:
pcounts[p] += (psum * norm_term)
return pcounts
def manysp_all_prots(sps, elutfs, **kwargs):
d_allprots = dict([(s,el.all_prots(elutfs, sp_base=s, **kwargs))
for s in sps])
return d_allprots
