Scale and transform Hi-C interaction data for visualization

Transforms and scales chromatin interaction data to prepare it for visualization. Applies user-defined scaling functions (e.g., log transformation) to interaction scores and handles missing values.

Usage

scaleData(data, scale_column, scale_method, remove_na = FALSE)

Arguments

data

Input data in one of these formats:

• ChromatinContacts object with imported interactions

• GInteractions object with score metadata

• data.frame or tibble with columns: seqnames1, start1, end1, seqnames2, start2, end2, plus score column

scale_column

Character. Name of column containing values to scale. Common options:

• "balanced": ICE-normalized counts (recommended)

• "count": raw contact counts

• Any other numeric metadata column

scale_method

Function to apply for transformation. Common options:

• log10: log10 transformation (default for most Hi-C data)

• log2: log2 transformation

• log1p: log(x + 1) transformation (handles zeros)

• identity or function(x) x: no transformation

• Custom function: any function that takes numeric vector and returns numeric vector

remove_na

Logical. Whether to remove rows with NA or infinite values after scaling (default: FALSE). Set TRUE to remove missing data that could cause visualization issues.

Value

Tibble (data frame) with standardized columns:

• seqnames1, start1, end1: First anchor coordinates

• seqnames2, start2, end2: Second anchor coordinates

• score: Transformed and scaled values

Details

Processing steps

1. Convert input to tibble format

2. Apply scale_method function to scale_column

3. Create new score column with transformed values

4. Optionally remove NA/infinite values

5. Squish extreme outliers to prevent visualization artifacts

Recommended scaling

For typical Hi-C data visualization:

• Use "balanced" column with log10 transformation

• Set remove_na = TRUE to handle bins with no coverage

Custom transformations

You can provide any function for scaling:

# Square root transformation
scaleData(cc, "count", sqrt)

# Custom transformation
scaleData(cc, "balanced", function(x) log2(x + 0.1))

See also

gghic(), geom_hic(), ChromatinContacts()

Examples

if (FALSE) { # \dontrun{
# Load Hi-C data
cc <- ChromatinContacts("file.cool") |> import()

# Standard log10 scaling of balanced data
scaled_data <- scaleData(cc, "balanced", log10)

# Raw counts without transformation
scaled_raw <- scaleData(cc, "count", function(x) x)

# Log2 scaling with NA removal
scaled_clean <- scaleData(cc, "balanced", log2, remove_na = TRUE)

# Use with plotting
library(ggplot2)
ggplot() +
  geom_hic(data = scaleData(cc, "balanced", log10),
           aes(seqnames1 = seqnames1, start1 = start1, end1 = end1,
               seqnames2 = seqnames2, start2 = start2, end2 = end2,
               fill = score))
} # }