deletions | additions       diff --git a/Notice_that_Seg7E1_t.tex b/Notice_that_Seg7E1_t.tex  index 7417737..9382b24 100644  --- a/Notice_that_Seg7E1_t.tex  +++ b/Notice_that_Seg7E1_t.tex 
...
Notice that Seg7E1 (top panel) appears to be more noisy than Seg7E2 (bottom panel). In the top panel, the red circle corresponds to genome site S7-503. Although this position is highlighted, $Ht(D_{S7-503})$ is close to the threshold, and there are other sites present higher values in both $Ht(D_i)$ and $Ht(N_i)$. Genome sites S7-91 and S7-112 also have above threshold $Ht(D_i)$ values, however, the $Ht(N_i)$ values at the two sites are even larger. Therefore, they do not belong to signal set.   In the bottom panel, there is only one green circle above the threshold, it is genome site S7-91. Similar to Seg7E1, its $Ht(N_{S7-91}) >Ht(D_{S7-91})$, it therefore does not belong to the signal set. Since there is not signal in Segment 7 based on Replicate II, we conclude that there is no substitution sites due to the treatment on this segment.  There is one site, S7-91, that consistently presents large $Ht(D_i)$ and $Ht(N_i)$ values across the two replicate. The proportions of its nucleotide read at each time point are shown in Figure \ref{fig:S7-91}. All four panels show complete transversion from G to C, with or without the treatment. The large values in $Ht(D_{S7-91})$ and $Ht(N_{S7-91})$ in both replicates precisely captures the read type switch that is not due to the treatment.