# Now, extract those gsites ...
dg_func = lambda x: pd.Series( ms.deamid_to_gsite(x['deamid_info'], x['start_fetched'], str(gbrecs[str(int(x['fetchid']))].seq)) )
# and add them back to the main table ...
gs_res = quant_n_raw[['deamid_info','start_fetched','fetchid']].apply( dg_func, axis=1 )
quant_n_raw = quant_n_raw.merge(gs_res,left_index=True,right_index=True)
print
print "Now we'd need to add theoretical glycosilation sites as a separate column ..."
print "full protein sequence and its length is added as well ..."
# this analysis must be done, once for each 'fetchid', and then merged back to the main table ...
get_theor_sites_fid = lambda fid: ms.get_theor_sites(str(gbrecs[str(fid)].seq))
get_theor_sites_number_fid = lambda fid: ms.get_theor_sites_number(str(gbrecs[str(fid)].seq))
theor_sites_info = lambda fid: pd.Series(
{'fetchid':fid,
'gsites_predicted':get_theor_sites_fid(fid),
'gsites_predicted_number':get_theor_sites_number_fid(fid),
'prot_seq':str(gbrecs[str(fid)].seq),
'prot_len':len(str(gbrecs[str(fid)].seq))} )
###################################################
predicted_gsite_info = quant_n_raw['fetchid'].drop_duplicates().apply(theor_sites_info)
# add back to the main table ...
quant_n_raw = quant_n_raw.merge(predicted_gsite_info,on='fetchid',how='right')
print "done ..."
print "numbering appears to be 1-based and overall correct!"
print
