trace 1.0.0

• Major Changes

  • Consolidated Workflow with trace() Function

    • Simplified three-step workflow:
      1. Read in data (e.g., read_fsa())
      2. Process with main trace() function
      3. Analyze with extraction functions (e.g., calculate_instability_metrics())
    • Deprecated functions (now internal-only):
      • add_metadata(), find_ladders(), find_fragments(), find_alleles(), call_repeats(), assign_index_peaks()
      • If used for legacy reasons access via the internal package with ‘:::’ (eg trace:::find_ladders()), these now require configuration via YAML file (see each functions help for more detail)
    • Configuration options:
      • Parameters can be passed via ... or YAML config file
      • YAML support enables reproducible analysis configurations and easier sharing of processing parameters

  • Enhanced Ladder Assignment Algorithm

    • New branching approach selects best overall fit rather than greedy assignment
      • Parameters:
        • ladder_top_n_branching: Controls number of branches evaluated at each step
        • ladder_branching_r_squared_threshold: Aggressiveness of branch pruning
        • ladder_assign_left_to_right: Direction of assignment (small→large or large→small)
    • Parameter name updates:
      • For ladder processing, minimum_peak_signal to minimum_ladder_signal
      • For ladder processing, scan_subset split into min_scan and max_scan

  • Improved Metrics Calculations

    • Enhanced quality filtering in calculate_instability_metrics():
      • New parameters: index_modal_signal_threshold, index_signal_sum_threshold
    • Fixed calculations:
      • Corrected weighting in skewness and kurtosis calculations
      • Fixed issue with average_repeat_change being non-zero for some of the index samples
• Data Handling Changes for Package Simplicity

  • Streamlined Data Import

    • Function updates:
      • repeat_table_to_repeats() to repeat_table_to_fragments() (now requires specific column names)
      • peak_table_to_fragments() split into:
        • genemapper_table_to_fragments() (for GeneMapper files)
        • size_table_to_fragments() (generic import)

  • Strict Metadata Requirements

    • Require exact columns (if those column names not detected, those metadata will be skipped):
      • unique_id, metrics_group_id, metrics_baseline_control , batch_run_id, batch_sample_id, batch_sample_modal_repeat
• Other Changes

  • Parameter Default Updates

    • Changed many defaults from NULL to NA to support configuration files

  • Fragment Detection Improvements

    • New peak_scan_ramp parameter in trace() (default relaxed from 6 to 5 to improve peak detection at longer sizes)

  • Internal Changes

    • find_fragments() now modifies in place (advanced usage only)

trace 0.5.0

• expansion_ratio metric updated so that it starts at 1 rather than 0

  • It now includes the relative signal of the index peak in the metric

• find_fragments filters minimum_peak_signal on the raw signal rather than the smoothed signal

• extract_alleles now will always return allele_2 so that the column names are always consistent

trace 0.4.0

Along with a new name, this is a major update to the previous version of the package. The updates below are changes relative to the main branch of the ‘instability’ github branch.

Breaking changes

• Most main functions in the pipeline modify in place rather than return a new object for efficient use of memory

  • This applies to the following functions: find_ladders(), add_metadata(), find_alleles(), call_repeats(), and assign_index_peaks().

• Repeat correction changes

  • The argument repeat_length_correction has been removed from call_repeats() and been replaced with correction
  • Now you chose between “repeat” or “batch” correction (see call_repeats() documentation for more info)

• Metadata categories renamed

  • Renamed for clarity and to make it clearer which functionality it’s associated with
  • plate_id -> batch_id, group_id -> metrics_group_id
  • size_standard_repeat_length -> batch_sample_modal_repeat
  • “size_standard” removed due to unnecessary redundancy of category

• Renamed instability metrics for clarity:

  • average_repeat_gain to average_repeat_change
  • modal_repeat_delta to modal_repeat_change

• Renamed number_of_peaks_to_return to number_of_alleles in find_alleles()

• Allele and index peak information can no longer be directly interacted with in the class

  • Setters (fragment_repeats$set_allele_peak() & fragment_repeats$set_index_peak()) and getters (fragment_repeats$get_allele_peak() & fragment_repeats$get_index_peak()) have been introduced so users don’t accidentally break things if they chose to directly interact with objects.

• Removed backwards compatibility of index assignment in calculate_instability_metrics()

  • grouped and index_override_dataframe parameters removed from calculate_instability_metrics()

• Renamed spike_location parameter in find_ladders() to ladder_start_scan

• Removed fix_ladders_auto() function

• The function generate_instability_template() was renamed to generate_trace_template()

• fragment_trace class is now instantiated after read_fsa() function rather than find_ladders()

• Throughout the package renamed any mention of height to signal for consistency

New features

• Batch correction

  • Correct systematic batch effects that occur across fragment analysis runs by using common samples across runs without having to know the repeat length

  • New associated function plot_batch_correction_samples() to help visualize this batch correction

  • See documentation for call_repeats() for more info

• New instability metric instability_index_change

• New function extract_ladder_summary() to get a summary of the ladder fits for all samples

• New function plot_data_channels() to help identify the channels containing the ladder and raw data

Bug fixes

• If some samples do not having a metrics_group_id or metrics_baseline_control the calculate_instability_metrics() function will still run but those samples will have missing values