4: The Tidyverse for Biological Data Wrangling

Chapter 4: The Tidyverse for Biological Data Wrangling

Genomic datasets frequently arrive as messy spreadsheets, variant tables, or multi-sample expression matrices. The Tidyverse (featuring dplyr, tidyr, and the pipe operator) offers an intuitive grammar for filtering, mutating, reshaping, and summarizing biological observations.

1. The Pipe Operator (%>% and |>)

Pipes pass the result of the left-hand expression as the first argument to the right-hand function, transforming deeply nested code into readable linear pipelines:

library(dplyr)

# Filter for significant cancer drivers and sort by magnitude
sig_drivers <- de_table %>%
    filter(padj < 0.01, abs(log2FoldChange) >= 1.5) %>%
    mutate(
        direction = ifelse(log2FoldChange > 0, "UP", "DOWN"),
        neg_log10_padj = -log10(padj)
    ) %>%
    arrange(desc(abs(log2FoldChange)))

2. Cohort Aggregations with group_by & summarise

Computing summary statistics across biological chromosome groups or experimental cohorts:

chr_summary <- sig_drivers %>%
    group_by(chr) %>%
    summarise(
        gene_count = n(),
        mean_fc    = mean(log2FoldChange),
        max_sig    = max(neg_log10_padj)
    )
print(chr_summary)

3. Reshaping Count Matrices: Wide to Long with pivot_longer

Standard count matrices are wide (one column per sample), but ggplot2 requires long tidy data:

library(tidyr)

long_counts <- wide_counts %>%
    pivot_longer(
        cols = -gene_id,
        names_to = "Sample_ID",
        values_to = "Read_Count"
    ) %>%
    mutate(Condition = ifelse(grepl("^Ctrl", Sample_ID), "Control", "Treated"))

print(head(long_counts, 6))