In npjc/mtpview1: Easily view microtitre plate data.

knitr::opts_chunk$set(echo = TRUE)

suppressPackageStartupMessages(library(tidyverse))
library(mtpview1)
mtp_example1

View your data in place interactively with mtp_view().

library(tidyverse)
# mtp_example1 %>% mtp_view()
# mtp_example1 %>% mtp_view(fillVar = "drug")
mtp_example1 %>% mtp_view(fillVar = "drug", fillOpacityVar = "conc")

Or, use mtp_plot() which is best for saving and printing and plays nice with ggplot2.

One plate

mtp_example1 %>% 
    mutate(plate = 1) %>% 
    mtp_ggplot(aes(plate = plate, well = well)) + 
    mtp_spec_96well() + 
    geom_footprint() + 
    geom_notched_border() +
    geom_col_label() + 
    geom_row_label() + 
    geom_well_rect(aes(fill = drug, alpha = conc))

add lineplots

You can also add lineplots to if you have repeated measures for each well:

library(tidyverse)
library(mtpview1)
mtp_example2 %>% 
    mutate(plate = 1) %>% 
    unnest() %>% 
    mtp_ggplot(aes(plate = plate, well = well)) + 
    mtp_spec_96well() + 
    geom_footprint() + 
    geom_notched_border() +
    geom_col_label() + 
    geom_row_label() + 
    geom_well_rect(aes(fill = cond)) + 
    geom_well_line(aes(x = runtime, y = measure))

Another example:

mtp_example3 %>% 
    mutate(plate = 1) %>% 
    unnest() %>% 
    mtp_ggplot(aes(plate = plate, well = well)) + 
    mtp_spec_96well() + 
    geom_footprint() + 
    geom_notched_border() +
    geom_col_label() + 
    geom_row_label() + 
    geom_well_rect(aes(fill = drug, alpha = dose_um)) + 
    geom_well_line(aes(x = runtime, y = measure)) + 
    geom_well_text(aes(label = strain))

Multiple Plates

data <- mtp_example4 %>% 
    filter(run == 1) %>% 
    unnest()

plot <- data %>% 
    mtp_ggplot(aes(plate = plate, well = well)) + 
    mtp_spec_96well() + 
    geom_footprint(aes(fill = strain), alpha = 0.25) +
    geom_notched_border() +
    geom_col_label(size = 2) +
    geom_row_label(size = 2) +
    geom_well_rect(fill = "grey96", size = 0.1) +
    geom_well_line(aes(x = runtime, y = measure), size = 0.2) + 
    facet_wrap(~plate)
plot

Roadmap

• [ ] unit testing and visual diff testing.
• [ ] cleaner interface that reduces duplication. Currently bandscales and placements are recomputed for each layer indepentently. This allows for max flexibility but this isn't necessarily worth it for the amount of complexity it introduces.
• [ ] dependencies: ggplot2, dplyr, tidyr, tibble... that's a lot...should attempt to minimize
• [ ] performance: eager, non-repetitive computer and straight to grid style?

npjc/mtpview1 documentation built on Oct. 29, 2019, 9:41 a.m.