CysingerR

CysingerR is an R package for analyzing deep sequencing data from cysteine scanning mutagenesis libraries. A tool that uses cysteine as a molecular probe to "sing out" which positions are on a protein's surface. It ports and enhances the original Perl pipeline from Najar et al. (Structure, 2017) into a single, easy-to-use R package with publication-quality visualizations.

Cysteine scanning mutagenesis systematically replaces each amino acid in a protein with cysteine, then uses thiol-reactive fluorescent probes and FACS sorting to identify surface-exposed vs. buried residues. This technique enables:

• 🎯 Conformational epitope mapping — identify where antibodies bind. The method can identify immunodominant regions on pathogen antigens — the regions that elicit the strongest antibody responses. This is directly useful for rational vaccine design, as it reveals which epitopes are preferentially targeted by the immune system.
• 💊 For antibody drug development - Knowing the precise epitope of a therapeutic antibody is essential for intellectual property protection, understanding mechanism of action, and predicting potential escape mutations. This method provides single-residue epitope mapping without requiring co-crystal structures, which are expensive and time-consuming to obtain. It can rapidly characterize panels of candidate antibodies and group them by epitope, facilitating selection of antibodies targeting distinct sites for combination therapies.
• 🏗️ Structural topology — determine surface accessibility without crystallography
• 🧬 Protein engineering — identify permissive sites for modifications

CysingerR provides the complete computational pipeline: from raw FASTQ reads to normalized enrichment ratios and publication-ready figures.

# Install from GitHub
# install.packages("devtools")
devtools::install_github("go2bean/CysingerR")

# Or install from a local source
install.packages("path/to/Cysinger_1.0.0.tar.gz", repos = NULL, type = "source")

Required (installed automatically):

• ggplot2 (≥ 3.4.0), dplyr, tidyr, patchwork, scales, rlang

Optional (for full pipeline functionality):

# Bioconductor packages for alignment
if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install(c("Biostrings", "ShortRead"))

library(Cysinger)

# Load built-in CcdB example data
cys_data <- load_example_data("residue_level")

# Normalize: calculate enrichment ratios
cys_norm <- normalize_cys_frequency(cys_data, input_bin = 1, sorted_bins = 2:4)

# Generate figures
plot_cys_heatmap(cys_data, bin_labels = c("Input", "Unsorted", "Low Sort", "High Sort"))
plot_epitope_profile(cys_norm, bin_labels = c("Unsorted", "Low Sort", "High Sort"))
plot_mutation_landscape(cys_data, bin_to_plot = 4, bin_label = "High Sort")

Cysinger generates 7 publication-quality visualizations with a premium dark theme:

Shows read counts across all protein positions (x-axis) and sorting bins (y-axis) using a log-scaled inferno color palette.

Loess-smoothed enrichment profiles across the protein sequence with highlighted structural regions. Peaks indicate surface-exposed positions; valleys indicate buried positions.

Lollipop plot showing read counts per position with automatic labeling of the most enriched positions.

Four-panel composite: total reads per bin, read count distributions, inter-bin correlation heatmap, and most variable positions.

Grouped bar chart showing raw read counts for each position, colored by sorting bin.

Polar plot classifying each position as surface-accessible (red) or buried (teal) based on enrichment.

Donut chart showing the proportion of WT, single-mutant, double, triple, and multi-mutant reads.

For processing raw FASTQ data end-to-end:

results <- run_pipeline(
  fwd_fastq  = "reads_forward.fastq",
  rev_fastq  = "reads_reverse.fastq",
  wt_seq     = ccdB_wt_nucleotide(),
  mid_tags   = c("ACGAGT", "ACGCTG", "AGACTG"),
  primers    = c("ATGCAG", "TTCAGG"),
  n_bins     = 4L,
  min_quality = 20L,
  min_length  = 75L
)

# Access results
results$cys_normalized       # Enrichment ratios
results$classification       # Mutation class summary
results$combine_stats        # Read merging QC

# Genetic code and sequence tools
get_codon_table()                        # Standard genetic code (64 codons)
translate_sequence("ATGTGTAAATGA")       # "MCK*"
reverse_complement_seq("ATGC")           # "GCAT"

# Wild-type CcdB reference
ccdB_wt_nucleotide()                     # 306 nt sequence
ccdB_wt_protein()                        # 102 aa (101 + stop)

The package includes CcdB cysteine scanning data from Najar et al., 2017:

# Residue-level cysteine frequencies (101 positions × 4 bins)
cys_data <- load_example_data("residue_level")

# Wild-type CcdB FASTA sequence
wt_seq <- load_example_data("wt_sequence")

# Pre-normalized dataset (ready for plotting)
cys_norm <- generate_example_normalized()

If you use Cysinger in your research, please cite:

Najar TA, Khare S, Suresh R, Pandey S, Raghavan V, Bhatt TK. Mapping protein binding sites and conformational epitopes using cysteine scanning and deep sequencing. Structure. 2017;25(3):395-405. doi: 10.1016/j.str.2017.06.015

MIT © TL Lee

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.