load_tcell fails on NaN 'EpitopeLinearSequence'

[arahuja]

File "/Users/arahuja/anaconda/lib/python2.7/site-packages/epitopes-0.1-py2.7.egg/epitopes/iedb.py", line 204, in load_tcell
verbose = verbose)
File "/Users/arahuja/anaconda/lib/python2.7/site-packages/epitopes-0.1-py2.7.egg/epitopes/iedb.py", line 79, in _load_dataframe
epitopes = epitopes.map(make_alphabet_transformer(reduced_alphabet))
File "/Users/arahuja/anaconda/lib/python2.7/site-packages/pandas/core/series.py", line 1980, in map
mapped = map_f(values, arg)
File "inference.pyx", line 920, in pandas.lib.map_infer (pandas/lib.c:43334)
File "/Users/arahuja/anaconda/lib/python2.7/site-packages/epitopes-0.1-py2.7.egg/epitopes/features.py", line 26, in transform
return ''.join([chr(48 + reduced_alphabet[char]) for char in s])
TypeError: 'float' object is not iterable

One of the sequences is loaded in as nan

[iskandr]

Arun, what args were you calling load_tcell with? I can't find the mystery NaN.

[arahuja]

iedb.load_tcell_ngrams(
assay_group='cytotoxicity',
human = True,
max_ngram = 2,
return_transformer = True,
reduced_alphabet = reduced_alphabet.murphy10)

[iskandr]

Thanks. I was transforming the strings before filtering out the null entries. I moved reduced_alphabet to the bottom of _load_dataframe and now it seems to work fine. Obviously needs more unit tests for all these filtering/transform options.