diff --git a/analysis/paper/plot/large_segments/plot_dnm_comparisons.R b/analysis/paper/plot/large_segments/plot_dnm_comparisons.R
index 26dbd68..5a8a02d 100755
--- a/analysis/paper/plot/large_segments/plot_dnm_comparisons.R
+++ b/analysis/paper/plot/large_segments/plot_dnm_comparisons.R
@@ -663,6 +663,7 @@ sapply(names(dnms), function(cohort){
 
 # Collapsed distribution of excess DNMs across all NDD segments
 pct.cutoff <- 0
+width.buffer <- 0
 sapply(names(dnms), function(cohort){
   if(cohort == "DDD_plus_ASC"){
     min.excess.for.norm <- 1.9
@@ -670,9 +671,11 @@ sapply(names(dnms), function(cohort){
   }else if(cohort == "DDD"){
     min.excess.for.norm <- 0.9
     min.excess.ineq <- ">="
+    width.buffer <- 0.3
   }else{
     min.excess.for.norm <- 0.1
     min.excess.ineq <- ">"
+    width.buffer <- 0.3
   }
   sapply(csqs, function(csq){
     cnv.marker.wex <- 0
@@ -701,7 +704,7 @@ sapply(names(dnms), function(cohort){
                                round(min.excess.for.norm, 0), ")'", sep="")
     }
     pdf(paste(out.prefix,  cohort, csq, "excess_dnm_distrib_bygene.pdf", sep="."),
-        height=2, width=2.75)
+        height=2, width=2.75+width.buffer)
     dnm.excess.cdf.barplots(ndd.segs[which(ndd.segs$cnv %in% cnv.types), ],
                             dnms[[cohort]], cohort=cohort, csq=csq,
                             norm=F, legend=T, pct.cutoff=pct.cutoff,
@@ -710,7 +713,7 @@ sapply(names(dnms), function(cohort){
     dev.off()
     cat(paste(cohort, csq, "\n"))
     pdf(paste(out.prefix, cohort, csq, "excess_dnm_distrib_bygene.norm.pdf", sep="."),
-        height=2, width=2.5)
+        height=2, width=2.5+width.buffer)
     dnm.excess.cdf.barplots(ndd.segs[which(ndd.segs$cnv %in% cnv.types), ],
                             dnms[[cohort]], cohort=cohort, csq=csq,
                             min.excess=min.excess.for.norm, norm=T, legend=F,
diff --git a/analysis/paper/plot/large_segments/plot_example_miami.R b/analysis/paper/plot/large_segments/plot_example_miami.R
index 6f75a6a..b040b16 100755
--- a/analysis/paper/plot/large_segments/plot_example_miami.R
+++ b/analysis/paper/plot/large_segments/plot_example_miami.R
@@ -178,6 +178,7 @@ mini.miami <- function(del, dup, max.p=10,
 ### RSCRIPT BLOCK ###
 #####################
 require(rCNV2, quietly=T)
+require(rasterpdf, quietly=T)
 require(optparse, quietly=T)
 
 # List of command-line options
@@ -231,4 +232,11 @@ png(out.png, height=dim.scalar*1.4*300, width=dim.scalar*1.95*300, res=300, fami
 mini.miami(del, dup, del.cutoff=del.cutoff, dup.cutoff=dup.cutoff, cutoff.label=cutoff.label,
            middle.axis.width=1.25, xaxis.label=xaxis.label, sig.col=graphabs.green, max.p=8)
 dev.off()
+out.pdf <- gsub(".png$", ".pdf", out.png)
+raster_pdf(out.pdf, height=dim.scalar*1.4, width=dim.scalar*1.95, units="in", res=600)
+# pdf(out.pdf, height=dim.scalar*1.4, width=dim.scalar*1.95)
+mini.miami(del, dup, del.cutoff=del.cutoff, dup.cutoff=dup.cutoff, cutoff.label=cutoff.label,
+           middle.axis.width=1.25, xaxis.label=xaxis.label, sig.col=graphabs.green, max.p=8)
+dev.off()
+
 
diff --git a/analysis/paper/plot/large_segments/plot_sliding_window_pval_distribs.R b/analysis/paper/plot/large_segments/plot_sliding_window_pval_distribs.R
index 71fd52c..485224e 100755
--- a/analysis/paper/plot/large_segments/plot_sliding_window_pval_distribs.R
+++ b/analysis/paper/plot/large_segments/plot_sliding_window_pval_distribs.R
@@ -319,6 +319,7 @@ primary.vs.secondary.scatter <- function(pvals.1, pvals.2, keep.idx=NULL,
 ### RSCRIPT BLOCK ###
 #####################
 require(rCNV2, quietly=T)
+require(rasterpdf, quietly=T)
 require(optparse, quietly=T)
 
 # List of command-line options
@@ -404,9 +405,14 @@ cat(paste("Median lambda, deletions & duplications only:", round(median(c(del.1$
 n.plots.wide <- 9
 n.plots.tall <- ceiling(length(hpos) / n.plots.wide)
 dim.scalar <- 7.5 / 10
-png(paste(out.prefix, "qq_grid_byHPO.png", sep="."),
-    height=dim.scalar*300*n.plots.tall, width=dim.scalar*300*n.plots.wide,
-    res=300)
+# png(paste(out.prefix, "qq_grid_byHPO.png", sep="."),
+#     height=dim.scalar*300*n.plots.tall, width=dim.scalar*300*n.plots.wide,
+#     res=300)
+raster_pdf(paste(out.prefix, "qq_grid_byHPO.pdf", sep="."),
+           height=dim.scalar*n.plots.tall, width=dim.scalar*n.plots.wide,
+           units="in", res=600)
+# pdf(paste(out.prefix, "qq_grid_byHPO.pdf", sep="."),
+#     height=dim.scalar*n.plots.tall, width=dim.scalar*n.plots.wide)
 par(mfrow=c(n.plots.tall, n.plots.wide))
 for(hpo in hpos){
   mini.qq(del.1, dup.1, hpo, title.cex=0.42, axis.cex=0.7, blue.bg=FALSE)
@@ -426,14 +432,22 @@ lambda.scatter(del.1, dup.1, hpo.n, blue.bg=FALSE)
 dev.off()
 
 # Scatterplots of primary vs. secondary P-values
-png(paste(out.prefix, "primary_vs_secondary_pvalue.DEL.png", sep="."),
-    height=prim.vs.sec.png.dim*300, width=prim.vs.sec.png.dim*300, res=300)
+# png(paste(out.prefix, "primary_vs_secondary_pvalue.DEL.png", sep="."),
+#     height=prim.vs.sec.png.dim*300, width=prim.vs.sec.png.dim*300, res=300)
+raster_pdf(paste(out.prefix, "primary_vs_secondary_pvalue.DEL.pdf", sep="."),
+           height=prim.vs.sec.png.dim, width=prim.vs.sec.png.dim, units="in", res=600)
+# pdf(paste(out.prefix, "primary_vs_secondary_pvalue.DEL.pdf", sep="."),
+#     height=prim.vs.sec.png.dim, width=prim.vs.sec.png.dim)
 primary.vs.secondary.scatter(del.1$pvals, del.2$pvals, cutoff.1=del.cutoff,
                              sig.color=cnv.colors[1], nonsig.color=control.cnv.colors[1],
                              pt.cex=0.175, blue.bg=FALSE, parmar=c(2.5, 2.5, 0.15, 0.15))
 dev.off()
-png(paste(out.prefix, "primary_vs_secondary_pvalue.DUP.png", sep="."),
-    height=prim.vs.sec.png.dim*300, width=prim.vs.sec.png.dim*300, res=300)
+# png(paste(out.prefix, "primary_vs_secondary_pvalue.DUP.png", sep="."),
+#     height=prim.vs.sec.png.dim*300, width=prim.vs.sec.png.dim*300, res=300)
+raster_pdf(paste(out.prefix, "primary_vs_secondary_pvalue.DUP.pdf", sep="."),
+           height=prim.vs.sec.png.dim, width=prim.vs.sec.png.dim, units="in", res=600)
+# pdf(paste(out.prefix, "primary_vs_secondary_pvalue.DUP.pdf", sep="."),
+#     height=prim.vs.sec.png.dim, width=prim.vs.sec.png.dim)
 primary.vs.secondary.scatter(dup.1$pvals, dup.2$pvals, cutoff.1=dup.cutoff,
                              sig.color=cnv.colors[2], nonsig.color=control.cnv.colors[2],
                              pt.cex=0.175, blue.bg=FALSE, parmar=c(2.5, 2.5, 0.15, 0.15))
diff --git a/docker/rcnv-base/Dockerfile b/docker/rcnv-base/Dockerfile
index 0b41676..120a34e 100644
--- a/docker/rcnv-base/Dockerfile
+++ b/docker/rcnv-base/Dockerfile
@@ -88,7 +88,7 @@ RUN apt-get -qqy clean && \
 # (Note that beeswarm _is_ available through conda but we need >=v0.4, and there 
 #  is a very complex dependency issue in conda that makes it easier to install 
 #  directly with R here)
-RUN Rscript -e "sapply(c('EQL', 'beeswarm'),\
+RUN Rscript -e "sapply(c('EQL', 'beeswarm', 'rasterpdf'),\
     function(pkg){install.packages(pkg,repos='https://cloud.r-project.org')})"
 ARG rCNV2_Rpkg_source="/opt/rCNV2/source/rCNV2_0.1.0.tar.gz"
 RUN Rscript -e "install.packages('$rCNV2_Rpkg_source', source=TRUE, repos=NULL)"