Minor changes to discussion chapter.

discussion/discussion.tex

@@ -120,7 +120,7 @@ \subsection{Implications of the two-pathway model in humans and dogs}
 
 The two pathway model categorizes crossovers into two classes: those that are subject to strong positive interference, and those with no interference.
 A suggestion from \citet{Housworth2003} was that these two pathways could correspond to crossovers placed by different mechanisms that are temporally separated.
-Non-interfering events were predicted to occur early in meiosis, and aided in the pairing and synapsis of homologues into the SC.
+Non-interfering events were predicted to occur early in meiosis, and aided in the pairing and synapsis of homologues into the synaptonemal complex (SC).
 Events that occurred later were part of a disjunction pathway, and subject to strong interference that spaced out the events along each chromosome.
 The wider spacing of crossovers likely assists in proper disjunction, reducing the risk of aneuploidy.
 %Further expansions of this model to fit data in yeast suggest that gene conversion is limited to the early pairing pathway, which is non-interfering\cite{Stahl2010}.
@@ -140,7 +140,7 @@ \subsection{Implications of the two-pathway model in humans and dogs}
 I found evidence for strong positive crossover interference in dogs, supporting previous cytological data\cite{Basheva2008}, and the two pathway model was favored over the simple gamma model.
 This adds to existing findings from this study and others\cite{Axelsson2012,Auton2013,Wong2010} that dog recombination is broadly similar to that of humans, apes, and mice.
 While the sample size of the dog study makes a firm conclusion difficult, 
-the support for the two pathway model of interference in dogs suggests that PRDM9 is likely not involved with interference.
+the support for the two pathway model of interference in dogs would suggest that PRDM9 is likely not involved with interference.
 This supports the idea that PRDM9 is part of an early acting, non-interfering recombination initiation pathway.
 
 
@@ -230,9 +230,9 @@ \section{Proposed model for recombination initiation and resolution}
 These regions are easily accessible by recombination initiation machinery that may provide an easy target for DSBs.
 
 In species that depend on it, PRDM9 is known to act on a fine scale to direct recombination into narrow hotspots.
-It appears that there are a proportion of crossovers in humans that may be independent of PRDM9, and this is supported by the finding of a healthy mother without PRDM9\cite{Narasimhan2016}.
+It is possible that there are a proportion of crossovers in humans that may be independent of PRDM9, and this is supported by the finding of a healthy mother without PRDM9\cite{Narasimhan2016}.
 In addition, hotspot overlap varies considerably on an individual basis\cite{Coop2008,Campbell2015}.
-Therefore it is likely that crossovers are placed with both PRDM9 dependent and independent mechanisms, but it is not known what proportion belong to each category.
+Therefore it is possible that crossovers are placed with both PRDM9 dependent and independent mechanisms, but it is not known what proportion belong to each category.
 An attractive explanation is to merge these crossover categories with those in the two pathway interference model.
 The support for the two pathway model of crossover interference in dogs (Chapter \ref{ch:dogPed}) in the absence of PRDM9 suggests that PRDM9 is not involved in interference.
 This therefore suggests that PRDM9 is part of an early crossover pathway that is independent of interference.