overlapping histograms - R

finally! after 2 days the result:

seemingly all go for ggplot functions.... even though i was a bit reluctant to use this function, tried for the past two days adjusting the df in order to apply ggplot... without success (yet ;-) )

however, i stumbled upon code for a lovely little pseudo graph at Chrismiller's blog (chrisamiller.com$BioBits) 

THANKS Chris!!

#i used for DNAmeth/comb.pair:

#at first to simply overlap a and b in hist:

a <- comb.pair="">

b <- comb.pair="">

hist( a, breaks= 40, col= "grey20" )

hist( b, breaks= 40, add= T ) #hahaha here is the trick: add= T!!

#so here is his code http://chrisamiller.com/science/2010/07/20/transparent-overlapping-histograms-in-r/

wtih a couple of other tricks on the same theme

#adapted for DNAmeth$comb.pair it looks like this:

plotOverlappingHist <- a="" b="" colors="c(" function="" gray20="" gray50="" span="" white="">

                                breaks=40, xlim=NULL, ylim=NULL){

  ahist=NULL

  bhist=NULL

  

  if(!(is.null(breaks))){

    ahist=hist(a,breaks=breaks,plot=F)

    bhist=hist(b,breaks=breaks,plot=F)

  } else {

    ahist=hist(a,plot=F)

    bhist=hist(b,plot=F)

    dist = ahist$breaks[2]-ahist$breaks[1]

    breaks = seq(min(ahist$breaks,bhist$breaks),max(ahist$breaks,bhist$breaks),dist)

    ahist=hist(a,breaks=breaks,plot=F)

    bhist=hist(b,breaks=breaks,plot=F)

  }

  if(is.null(xlim)){

    xlim = c(min(ahist$breaks,bhist$breaks),max(ahist$breaks,bhist$breaks))   

  }

  if(is.null(ylim)){

    ylim = c(0,max(ahist$counts,bhist$counts))   

  }

  overlap = ahist

  for(i in 1:length(overlap$counts)){  

    if(ahist$counts[i] > 0 & bhist$counts[i] > 0){

      overlap$counts[i] = min(ahist$counts[i],bhist$counts[i])

    } else {

      overlap$counts[i] = 0

    }

  }

  plot(ahist, xlim=xlim, ylim=ylim, col=colors[1])

  plot(bhist, xlim=xlim, ylim=ylim, col=colors[2], add=T)

  plot(overlap, xlim=xlim, ylim=ylim, col=colors[3], add=T) 

}

# and then he plots:

plotOverlappingHist(a,b)

#alternatively, someone suggested to use transparency.... riiight here we have the real solution and can be adapted for BGW or COL depending on publication preferences :-)

hist( am, col= rgb( 0, 1, 0,0.5 ), breaks= 40, main= Histogram of frequencies in Monocytes  )

# sit legend in: legend( "topleft", legend... hist( bm, col= rgb( 1, 0, 0,0.5 ), breaks= 40, add=T )

#bellissimo!!

#cell specific:

par( mfrow= c(1,2 ) )

a <- comb.pair="" font="">

am <- a="" font="" grep="" rownames="">

amhhc <- all="" am="" contain="" controls="" for="" from="" hhc:="" household="" m="" span="">

amtb <- active="" all="" am="" cases="" contain="" for="" from="" m="" span="" tb:="" tb="">

hist( amhhc, col= rgb( 0, 1, 0,0.5 ), breaks= 40, main= Histogram of frequencies in Monocytes  )

hist( amtb, col= rgb( 1, 0, 0,0.5 ), breaks= 40, add=T )

an <- a="" font="" grep="" rownames="">

anhhc <- all="" an="" contain="" controls="" for="" from="" hhc:="" household="" n="" span="">

antb <- active="" all="" an="" cases="" contain="" for="" from="" n="" span="" tb:="" tb="">

hist( anhhc, col= rgb( 0, 1, 0,0.5 ), breaks= 40, main= "Histogram of frequencies in Neutrophils")

hist( antb, col= rgb( 1, 0, 0,0.5 ), breaks= 40, add=T )

***************************************
op die ouend wat goed gewerk het op een figuur:
--------------
par( mfrow = c( 1,2 ) )
hist( amhhc, col= rgb( 0,0,0,0.5 ), breaks= 60, main= "Histogram of frequencies in Monocytes", xlab="" )
hist( amtb, col= rgb( 1,1,1,0.5 ), breaks= 60, add= T, main= "", xlab= "" )
hist( anhhc, col= rgb( 0,0,0,0.5 ), breaks= 60, main= "Histogram of frequencies in Neutrophils", xlab="" )

hist( antb, col= rgb( 1,1,1,0.5 ), breaks= 60, add= T, main= "", xlab= "" )

----------------

and then of course the entertainment - ooohhh how we know these!!

http://chrisamiller.com/science/2011/01/21/bad-project/