Add annotation_col in pheatmap

Question

[eimanpharmacist]

Hi, I am working in microbiome analysis and am new in using pheatmap package. I have approx 160 samples from 2 field trials (2 years) with two different treatments (control and pathogen_treated). I calculated the relative abundance of the top50 OTUs for each tissue treatment (for example, top50 OTUs of 40 healthy tissues in 2017, ... same for each treatment per each year), Then I summed the RA of each OTU for all samples in the same year and of the same treatment, so 160 samples are now represented in 4 sets (healthy2016, pathogen2016, healthy2017, pathogen2017) and since the top50 are relatively different in each set, so finally I have a table summed RA of 91 OTUs across 4 sets of samples.

    Healthy_2016    pathogen_2016   Healthy_2017    pathogen_2017
K_Bacteria (OTU92)  0.0836997   0   0   0
Enterococcus (OTU65)    0   0   0.4952662   0.0990601
Leuconostoc (OTU37) 0.1895487   0.1126888   0   0
Lactococcus (OTU106)    0   0   0.0425752   0
Lactococcus (OTU36) 0.2293602   0   1.0179649   0.7535034
Exiguobacterium (OTU38) 0   0   0.261444    0
Sphingobacterium (OTU71)    0.103809    0   0.0763031   0.1448188

I created a heatmap using the following code:

```{r fig.height=13, fig.width=8}
# create heatmap using pheatmap
pheatmap(dat, color = colorRampPalette(rev(brewer.pal(n = 7, name =
  "RdYlBu")))(100), kmeans_k = NA, breaks = NA, border_color = "grey60",
  cellwidth = NA, cellheight = NA, scale = "row", cluster_rows = TRUE,
  cluster_cols = TRUE, clustering_distance_rows = "euclidean",
  clustering_distance_cols = "euclidean", clustering_method = "complete", cutree_rows = NA, cutree_cols = 2, legend = TRUE, legend_breaks = -2:2,legend_labels = c("-1.5", "-1", "0", "1", "1.5"))

Then I need to add 2 annotation_col to the heatmap, first based on treatment so column 1 and 3 will be represented as healthy tissues and 2, 4 as pathogen-treated tissues, second annotation bar should represent the year of the field-trial, so columns 1 and 2 will be labeled fieldtrial2016 and 3 and 4 labeled fieldtrial2017. I checked different solutions online but none worked. So, I believe what I need is to create a dataframe of my-metadata to use it in col annotation. For example

Samples             Treatment           Year
Healthy_2016        HealthyTissue   FieldTrial2016
pathogen_2016   FG_Tissue            FieldTrial2016
Healthy_2017        HealthyTissue   FieldTrial2017
pathogen_2017   FG_Tissue           FieldTrial2017

I do not know how to create a dataframe of these repeated data to use in annotation_col

Your help is much appreciated

Answer 1

[Kevin Blighe]

Please take a look at my answer here: A: pheatmap annotation - legend only for columns

In my answer, which is reproducible, I show how to add both row and col annotations. Essentially, the rownames of the object that you pass to annotation_col and annotation_row has to match the colnames and rownames, respectively, of your input data-matrix.

Kevin