ex_vivos.rmd

---
title: "R Notebook"
output: html_notebook
---

Read the metabolomics data
```{r}
library(data.table)
library(readxl)
library(magrittr)

abundance <- read_excel("../ex_vivos_2019/metabolon_quant.xlsx", sheet="abundance", col_names = FALSE)
mets <- read_excel("../ex_vivos_2019/metabolon_quant.xlsx", sheet="metabolites") %>% setDT()
samples <- read_excel("../ex_vivos_2019/metabolon_quant.xlsx", sheet="samples") %>% setDT()

colnames(abundance) <- samples$sample_id
abundance$metabolite_id <- mets$compound_id
abundance <- as.data.table(abundance) %>% melt(id.vars = "metabolite_id", 
                                               variable.name = "sample_id", value.name = "abundance")
abundance <- mets[abundance, on = c(compound_id = "metabolite_id")]
abundance <- samples[abundance, on = "sample_id"]
```

Now we read the inferred composition:

```{r}
diet <- fread("data/mgx/food_contents.csv")
compounds <- diet[source_type == "Compound"]
```

```{r, fig.width=20, fig.height=16}
library(pheatmap)

food <- fread("data/mgx/food_abundances.csv")
means <- food[, .(reads = mean(reads), relative = mean(fraction)), by="taxon"]
M <- dcast(food, taxon ~ sample_id, value.var = "fraction", fun.aggregate = mean, fill = 0)
n <- M[, taxon]
M <- as.matrix(M[, taxon := NULL])
rownames(M) <- n
pheatmap(M, cluster_cols=FALSE)
```