R and Tidyverse Fundamentals

1. Objects and data frames

R stores columns as vectors. A data frame should have meaningful column names and explicit types.

samples <- data.frame(
  sample_id = c("S01", "S02", "S03"),
  condition = factor(c("control", "treated", "treated")),
  reads = c(1200000, 1450000, 1310000)
)
str(samples)

2. Tidy transformations

The tidyverse makes transformations readable. Always inspect the result after filtering or joining.

install.packages(c("tidyverse", "here"))
library(tidyverse)
samples |>
  mutate(millions = reads / 1e6) |>
  group_by(condition) |>
  summarise(mean_reads = mean(millions), .groups = "drop")

3. Visualize a QC metric

Plots should show units, labels, and the biological grouping used in the analysis.

ggplot(samples, aes(condition, millions, color = condition)) +
  geom_point(size = 3) +
  labs(x = "Condition", y = "Reads (millions)") +
  theme_minimal()
ggsave("results/read_qc.webp", width = 6, height = 4, dpi = 150)

4. Missing values and joins

Use is.na() to identify missing data and verify identifiers before joining metadata to measurements. A silent many-to-many join can duplicate observations.

samples |> summarise(across(everything(), ~sum(is.na(.x))))
stopifnot(!anyDuplicated(samples$sample_id))

5. Reading and Writing Data

In bioinformatics, you will frequently import CSV, TSV, and Excel files. Use the readr package (part of tidyverse) for consistent parsing and explicit column type detection.

# Read a tab-separated gene expression matrix
expr <- read_tsv("data/raw/gene_counts.tsv", show_col_types = TRUE)

# Inspect dimensions and first rows
dim(expr)
head(expr)

# Write results with consistent formatting
write_csv(expr, "data/processed/filtered_counts.csv")

Common pitfall: Excel silently converts gene names like MARCH1 and SEPT2 to dates. Always use plain-text formats (CSV/TSV) for bioinformatics data and verify gene name integrity after import.

6. Reshaping Data: Wide vs. Long Format

Many bioinformatics datasets arrive in wide format (one column per sample), but most R analysis and plotting functions expect long (tidy) format.

# Wide format: each sample is a column
wide_data <- data.frame(
  gene = c("TP53", "BRCA1", "MYC"),
  sample_A = c(150, 320, 890),
  sample_B = c(175, 280, 920)
)

# Pivot to long format for ggplot and statistical testing
long_data <- wide_data |>
  pivot_longer(
    cols = starts_with("sample_"),
    names_to = "sample",
    values_to = "expression"
  )

# Now each row is one gene-sample measurement
print(long_data)
# Output:
# gene   sample    expression
# TP53   sample_A  150
# TP53   sample_B  175
# BRCA1  sample_A  320
# ...

7. Factors and Categorical Variables

Factors are essential for controlling group order in plots and statistical models. Always set factor levels explicitly rather than relying on alphabetical order.

# Without explicit levels, R uses alphabetical order
samples$condition <- factor(samples$condition, levels = c("control", "treated"))

# This ensures "control" appears first in plots and is the reference level in models
levels(samples$condition)
# [1] "control" "treated"

Why this matters: In differential expression analysis, the reference level of a factor determines the direction of fold-change calculations. If "treated" is accidentally set as the reference, all log2FC values will be inverted.

8. Error Handling and Defensive Programming

Before running an analysis pipeline, validate your inputs programmatically:

# Check that required columns exist
stopifnot(
  "sample_id" %in% colnames(samples),
  "condition" %in% colnames(samples),
  nrow(samples) > 0
)

# Check for unexpected NA values in critical columns
na_counts <- samples |> summarise(across(everything(), ~sum(is.na(.x))))
print(na_counts)

# Assert no duplicate sample IDs
if (anyDuplicated(samples$sample_id)) {
  stop("Duplicate sample IDs detected – check your metadata file.")
}

9. Reproducible R Sessions

Always record your R session information at the end of every analysis script. This captures the exact R version, loaded packages, and operating system.

# At the end of your script
sink("logs/session_info.txt")
sessionInfo()
sink()

# Expected output includes:
# R version 4.4.1 (2024-06-14)
# Platform: x86_64-pc-linux-gnu
# attached packages: tidyverse 2.0.0, ggplot2 3.5.1, dplyr 1.1.4 ...

Practical Exercise

Create a table with sample ID, condition, and one QC metric. Produce one labeled plot, save it to results/, and write one sentence interpreting the pattern without overstating it.

Pass criteria: The table has correct types, the plot has units and labels, and the learner can explain how missing values and duplicate IDs would affect analysis.

Troubleshooting

If a package will not install, record the R version and use a project library. If a plot is empty, inspect factor levels and missing values before changing the code.

Next Steps

Continue with Statistics for Bioinformatics and Data Visualization Fundamentals. Record the software versions, dataset or example inputs, and any decisions you made.