Detection of outliers (peaks & valleys) in ONT signal using z-scores.

This is the training-set variant of the z-score-based peak/valley detection function. It identifies deviations in the poly(A) tail signal that may correspond to non-adenosine nucleotides (C, G, or U) by applying a moving-average z-score filter with empirically calibrated parameters.

Usage

filter_signal_by_threshold_trainingset(signal)

Arguments

signal

Numeric vector. An ONT read fragment corresponding to the poly(A) tail region as delimited by nanopolish polya (stored in tail_feature_list[[1]] produced by create_tail_feature_list_trainingset).

Value

Numeric vector of pseudomoves with values in {-1, 0, 1}:

1

Signal exceeds baseline (peak) — potential G nucleotide.

-1

Signal falls below baseline (valley) — potential C or U nucleotide.

0

Signal within baseline range — homopolymer A region.

Details

The algorithm prepends 100 calibration data points (sampled from the 10 most frequent signal values and the first 10 observations) to stabilise the baseline estimate. It then applies a sliding window of 100 data points: any observation exceeding 3.5 standard deviations from the running mean is classified as a peak (+1) or valley (-1); all other positions are scored 0. After trimming the calibration prefix, terminal positions are zeroed out (first 5 and last 5) to prevent edge artefacts from entering downstream chunk extraction. Finally, substitute_gaps is applied to fill isolated zero-gaps in the pseudomove vector.

Acknowledgements

The z-score peak detection algorithm is based on: Brakel, J.P.G. van (2014). "Robust peak detection algorithm using z-scores". Stack Overflow. Available at: https://stackoverflow.com/questions/22583391/peak-signal-detection-in-realtime-timeseries-data/22640362#22640362 (version: 2020-11-08).

See also

extract_tail_data_trainingset which calls this function, filter_signal_by_threshold for the production counterpart, substitute_gaps for the gap-filling post-processing step.

Examples

if (FALSE) { # \dontrun{

filter_signal_by_threshold_trainingset(
  signal = tail_feature_list[[1]][["readname"]][[4]])

} # }