Hi Charles,

Yes I am. Unfortunately I had already come across the page you linked and really what I'm trying to do is something like this image I have attached from this paper

"Influence of hydraulic regimes on bacterial community structure and composition in an experimental drinking water distribution system" I. Douterelo*, R.L. Sharpe, J.B. Boxall.

I just want to have the colour key showing the percentage scale.

Hi Sebastian,

Yes you see I really don't want to cluster them yet. If you think the chapter could help me then yes I would really appreciate it. I'm very new to this kind of analysis.

Thanks,

Ciara

Hi Sebastian,

Yes that looks perfect I'll try it today! I had started to follow this method http://www.molecularecologist.com/2013/08/making-heatmaps-with-r-for-microbiome-analysis/ which I found worked really well, alas I managed to get the dendrograms and colour plot scale but ended up removing my sample names (because it gets you to select the matrix data however in the example the authors example has the sample id's after this step but I can't seem to get that no matter what I tried and they only come up as numbered now). I guess I just need to understand the terms but I found following that guide helped, do you think the function above would work if I'm following the steps in this method?

Hi Metthijs,

I will try that package, although the heatmap I produced with the above post looks very good so I am happy with it apart from the new issue with sample names.

Thanks everyone for all the help I'm aware my skills in this kind of analysis are not the best but I'm learning.

Ciara

Hi Sebastian,

When I do that this error message comes up?

Error in UseMethod("levelplot") :

no applicable method for 'levelplot' applied to an object of class "data.frame"

>

Hi Sebastian,

I've gotten the heat maps to look exactly like the last one you uploaded there I'm really happy with it thanks for all the help. However, as I was doing the analysis as in that description I uploaded in my previous response. Because I tell it what the data matrix is without the sample ID's

ie.

>dim (all.data)

>row.names(all.data) all.data data.prop

Hi Sebastian,

Apologies but I have been back in the lab and very busy so I didn't get to get back to my R plots. Thank you so much for this I'm going to look at yours and try again with my own file. I did a quick check of mine and the code I was using and I am still losing the sample ID's and when I try to re tell it the row names it doesn't seem to change. I'll look at your example now. Thanks again.

Hi Sebastian,

So I input the following;

> source("/Users/ciarakeating2/Documents/GenusR.csv")

Error in source("/Users/ciarakeating2/Documents/GenusR.csv") :

/Users/ciarakeating2/Documents/GenusR.csv:1:7: unexpected ','

1: Sample,

^

> genus

> dim (genus)

[1] 14 397

> genus [1:3, 1:4]

Sample tm7..candidate.division. paenibacillus dongia

1 Seed DNA 1 0 0

2 Seed cDNA 0 0 0

3 R2 PMT DNA 1 0 0

> row.names(genus) genus data.prop data.prop [1:3, 1:3]

tm7..candidate.division. paenibacillus dongia

1 4.974382e-05 0 0

2 0.000000e+00 0 0

3 5.507518e-05 0 0

> heatmap(as.matrix(data.prop), Rowv = NA, Colv = NA, col = scaleyellowred)

> maxab head(maxab)

tm7..candidate.division. paenibacillus dongia

2.436796e-04 7.208246e-05 2.036798e-04

treponema proteocatella hydrogenophaga

7.651841e-04 9.618230e-04 8.545270e-03

> n1 data.prop.1 heatmap(as.matrix(data.prop.1), Rowv = NA, Colv = NA, col = scaleyellowred, margins = c(12, 2))

> data.dist row.clus heatmap(as.matrix(data.prop.1), Rowv = as.dendrogram(row.clus), Colv = NA, col = scaleyellowred, margins = c(10, 3))

> data.dist.g col.clus heatmap(as.matrix(data.prop.1), Rowv = as.dendrogram(row.clus), Colv = as.dendrogram(col.clus), col = scaleyellowred, margins = c(10, 3))

> heatmap.2(as.matrix(data.prop.1), Rowv = as.dendrogram(row.clus), Colv = as.dendrogram(col.clus), col = scaleyellowred, margins = c(11, 5), trace = "none", density.info = "none", xlab = "genera", ylab = "Samples", main = "Heatmap Genera", lhei = c(2, 8))

When I try input the code you have in your image I get this response

> rnames

Ah yes you are right! In previous attempts I had included the sep term...

Ok I will try this again and let you know I will attempt with everything you have said there and I'll make sure my csv file looks the same. Hopefully it will work then! I appreciate all the help. Thank you.

Hi Sebastian,

My .csv file looked like this (Family), and when I import it that way I can see the table by pressing the following;

Family dim(all.data)

[1] 14 41

> Family [1:3, 1:4]

Sample desulfuromonadaceae alteromonadaceae fibrobacteraceae

1 Seed DNA 3 1 0

2 Seed cDNA 8 2 0

3 R2 PMT DNA 9 4 0

when I changed it to yours (Family1) and used the command

Family

This is my table .csv file

And this is when I changed it to be like yours. Is it at this point that I'm wondering should I be using the read table function instead? See I'm following that above tutorial mentioned as it's been the only way I've managed to work with the matrix, calculate abundance and then the bray curtis similarity between samples and species to produce the dendrograms. As I've found other functions I've used have not given the correct result. So I'm just wondering is there a simpler way to modify what I'm doing so as to re tell R to put the row names back in, the row.names function doesn't seem to have any effect on the heat map.

Hi Sebastian,

I tried doing exactly how you described and the result was this.... Which is why it is important for me to do the abundance, eliminate lower %'s, and then do the clustering from there.

I'll try integrate your method with mine above and see how that goes. Or if you have any idea on how I can work from the data.prop stage in my answer above? As I need that command for further commands.

My file family in the .csv format comes up like this;

Sample,desulfuromonadaceae,alteromonadaceae,fibrobacteraceae,burkholderiaceae,rhodocyclaceae,gracilibacteraceae,cytophagaceae,legionellaceae,hyphomicrobiaceae,xanthobacteraceae,holophagaceae,op11 (candidate division) ,flammeovirgaceae,rhodobacteraceae,marinilabiaceae,actinosynnemataceae,thermodesulfobiaceae,pasteurellaceae,caldicellulosiruptoraceae,caulobacteraceae,peptococcaceae,flexibacteraceae,dermatophilaceae,bradyrhizobiaceae,ignavibacteriaceae,gn06 (candidate division) ,gn02 (candidate

The file I saved like your .csv file comes up like this one;

# heat map bacterial family data set,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

# 10/21/13 sr,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,desulfuromonadaceae,alteromonadaceae,fibrobacteraceae,burkholderiaceae,rhodocyclaceae,gracilibacteraceae,cytophagaceae,legionellaceae,hyphomicrobiaceae,xanthobacteraceae,holophagaceae,op11 (candidate division) ,

Can I ask why you have yours formatted this way are these headings?

Ah ok thank you! Is there a way in R after I calculate everything and have my data.prop as above answers am I able to use the heat map function and say what the sample ID's are? It's just when I use this function;

row.names(all.data)

Sebastian finally! Finished heatmap with sample ID's! I'm not sure why but I cleared the R console, re saved the file table, input row names and carried out the analysis with the matrix and somehow this time it worked out. Thanks for all your help, and the book looks great, I've to pass this knowledge onto new students in our lab who will be doing this analysis so that will be a go to guide!

Hi Clara,

The figure in your reference paper:

"Influence of hydraulic regimes on bacterial community structure and composition in an experimental drinking water distribution system" I. Douterelo*, R.L. Sharpe, J.B. Boxall.

was made with JColorGrid:

http://sourceforge.net/projects/jcolorgrid/

this is a platform independent program with a graphical user interface that was designed to make reasonable heatmaps with minimal effort. It won't create a dendrogram but you can provide an ordered matrix as input. R is great, but sometimes you just need a standard plot instead of cleaning input files and searching and testing for the right parameters.

best,

marcin