Derived.Rmd

---
title: "Derived"
author: "Brian Yandell"
date: "2023-10-03"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

I have an idea how to incorporate derived traits into the tool. Basically, we create a spreadsheet (CSV or XLSX) with columns for:
Left      Operator         Right    Derived
Glucose           -           lactate glu_over_lac
Lactate             -           pyr       lac_over_pyr
Glc_M1            +          Glc_M2            Glc_M1-2
 
We might want some special way to combine multiple traits (Say Glc M1-5).
 
In addition, I want to set this up so that we can operate on GTT measurements, such as
Glc_13C_M3_0_18wk to Glc_13C_M3_120_18wk together are known as Glc_13C_M3_18wk,
So we could do Glc_13C_M1_18wk + Glc_13C_M2_18wk to get the sum of these two as a new trait over time, Glc_13C_M1-2_18wk.

```{r}
library(dplyr)
library(tidyr)
library(readr)
```

```{r}
dirpath <- file.path("~", "founder_diet_study")
dirpath <- file.path(dirpath, "HarmonizedData")
traitSignal <- readRDS(file.path(dirpath, "traitSignal.rds")) %>%
  filter(dataset == "LivMet")
```

```{r}
derived <- read_csv("data/derived.csv") %>%
  pivot_longer(any_of(c("Left","Right")), names_to = "Side", values_to = "datatrait")
```

```{r}
derivedSignal <- 
  left_join(
    derived,
    traitSignal %>%
      unite(datatrait, dataset, trait, sep = ": "),
    by = "datatrait",
    relationship = "many-to-many")
```


```{r}
derivedSignal %>%
  select(-datatrait, -signal) %>%
  pivot_wider(names_from = "Side", values_from = "cellmean") %>%
  arrange(strain, sex, condition, Derived) %>%
  mutate(newtrait = Left - Right,
         newtrait = ifelse(Operator == "+", Left + Right, newtrait),
         newtrait = ifelse(Operator == "/", Left / Right, newtrait),
         newtrait = ifelse(Operator == "*", Left * Right, newtrait)) %>%
  select(-Operator, -Left, -Right) %>%
  rename(trait = "Derived",
         value = "newtrait") %>%
  select(strain, sex, condition, trait, value)
```


```{r}
rightSignal <- 
  left_join(
    derived %>%
      select(Right, Derived) %>%
      rename(datatrait = "Right"),
    traitSignal %>%
      unite(datatrait, dataset, trait, sep = ": "),
    by = "datatrait")
```

```{r}
full_join(leftSignal, rightSignal,
          by = c("Derived", "strain", "sex", ""))
```