Fixes and clarifications to doc homepage

doc/kwip.rst

@@ -84,9 +84,9 @@ of ``$HOME`` with your preferred installation prefix.
 
 
 The ``kwip`` executable is the core of ``kWIP``; its help statement is
-reproduced above. This program operates on the saved Count-min Sketches of
-``khmer`` (which they call Countgraphs). One can run with or without the
-population entropy weighting, using the ``-U`` parameter to disable weighting.
+reproduced above. This program operates on the saved Countgraphs of ``khmer``.
+One can run with or without the entropy weighting, using the ``-U`` parameter
+to disable weighting.
 
 An example command could be:
 
@@ -111,10 +111,9 @@ The Concepts Behind ``kWIP``
 The inner product between two vectors is directly related to the distance
 between the vectors in Euclidean space. This has been utilised several times in
 Bioinformatics to implement measures of genetic similarity between two
-sequences, including the :math:`D2` statistic, and the spectrum kernel for
-protein sequence compassion. Traditionally, the software which implement these
-and similar algorithms operate on known genetic sequences, e.g. those taken
-from a reference genome. ``kWIP``'s innovation is to weight the inner product
-operation by a weight vector, and to derive weights in a way which minimises
-the noise generated by next-gen sequencing protocols while maximising the
-signal of genetic distance between samples.
+sequences, including the :math:`D2` statistic. Traditionally, the software
+which implement these and similar algorithms operate on known genetic
+sequences, e.g. those taken from a reference genome. ``kWIP``'s innovation is
+to weight the inner product operation by a weight vector, and to derive weights
+in a way which minimises the noise inherent in next-gen sequencing datasets
+while maximising the signal of genetic distance between samples.