Sequence.gc methods that consider IUPAC nucleotide ambiguity

[mdshw5]

The existing Sequence.gc method purposefully ignores characters other than G/C and uses the sequence length as a denominator to produce "fraction g/c". This has a few benefits:

• we can ignore IUPAC ambiguous DNA codes
• len(sequence) is fast to compute vs. counting more occurrences of characters

The downside is that any non-GCAT characters may be included in the denominator:

pyfaidx/pyfaidx/__init__.py

Lines 254 to 266 in 7b4d8d7

	@property
	def gc(self):
	""" Return the GC content of seq as a float
	>>> x = Sequence(name='chr1', seq='ATCGTA')
	>>> y = round(x.gc, 2)
	>>> y == 0.33
	True
	"""
	g = self.seq.count('G')
	g += self.seq.count('g')
	c = self.seq.count('C')
	c += self.seq.count('c')
	return (g + c) / len(self.seq)

I'd welcome any pull request to implement something like:

• Sequence.gc_iupac method that counts e.g. S=GC and W=AT, and also considers K=GT. This is considerably more difficult than the current method and requires some validation of the sequence to confirm that it only contains valid IUPAC letters
• Sequence.gc_strict method that counts G/C and A/T, implicitly ignoring all other characters. This is probably closest to what people expect as GC content