Sat 206 indenpendent chr estimations by Cristianetaniguti · Pull Request #17 · Breeding-Insight/Qploidy2

Cristianetaniguti · 2026-08-21T19:28:17Z

No description provided.

Copilot

Pull request overview

This PR updates Qploidy2’s HMM copy-number pipeline to be more robust to chromosome-order effects and z-only (all-homozygous) windows, adds new quality/summary utilities for multi-sample results, and aligns plotting + documentation around the new CN_reliability metric and updated heterozygosity defaults.

Changes:

Refactors HMM decoding to use per-chromosome forward–backward smoothing and introduces CN_reliability (plus related plotting updates).
Adds multi-sample utilities (compare_cn_track_summary, filter_hmm_CN, count_types + print method) and expands documentation/tutorials accordingly.
Adjusts default heterozygosity gating (min_het_frac, het_range) across BAF/HMM/standardization workflows and updates tests.

Reviewed changes

Copilot reviewed 23 out of 34 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
tests/testthat/test-hmm_cn.R	Updates expected numeric outputs from HMM-related tests.
R/standardization.R	Tweaks re-standardization defaults and warns/overrides mismatched `ploidy.standardization`.
R/plots_standardization.R	Updates ratio labeling to θ and improves facet labeling.
R/plots_hmm_results_all_samples.R	Adds/relocates multi-sample CN plotting + summary plotting utilities.
R/plot_raw.R	Updates ratio plot y-axis label and raw-model defaults for het gating.
R/plot_karyotype.R	Adds `CN_reliability` coloring and better handling of filtered/NA CN segments.
R/plot_cn_track.R	Colors CN track by `CN_reliability` (fallback to legacy prob).
R/karyotype_notation.R	Suppresses NSE NOTE for `CN_call_filled`.
R/hmm_utils.R	Adds `fb_smooth()` and `viterbi_bidi()` utilities (internal).
R/hmm_main.R	Major HMM stability updates, new parameters (`z_no_baf_scale`, `hom_z_sigma_inflate`), adds `CN_reliability`, changes defaults.
R/filter_hmm_CN.R	New filtering + sample-type classification utilities (`filter_hmm_CN`, `count_types`, print method).
R/export_results.R	Adds `CN_mode` to globalVariables to address R CMD check NOTES.
R/em_hmm_cn.R	Adds `update_pi0` flag and reintroduces a mu-update stability guard.
R/call_dosages.R	Updates het gating defaults in dosage calling interface/docs.
R/BAF_distributions.R	Updates het gating defaults in BAF likelihood/model selection.
NEWS.md	Adds release notes for 1.18.x including new features and default changes.
NAMESPACE	Exports new functions and adds new imports/S3 registration.
man/select_best_raw_model.Rd	Updates documented defaults for het gating.
man/select_best_baf_model.Rd	Updates documented defaults for het gating.
man/re_standardize.Rd	Updates documented defaults for het gating.
man/print.count_types.Rd	New Rd for `print.count_types`.
man/plot_karyotype.Rd	Documents new `CN_reliability` option.
man/hmm_estimate_CN.Rd	Documents new parameters and default changes.
man/filter_hmm_CN.Rd	New Rd for `filter_hmm_CN`.
man/em_hmm_cn.Rd	Documents `update_pi0`.
man/count_types.Rd	New Rd for `count_types`.
man/compute_baf_likelihoods.Rd	Updates documented defaults for het gating.
man/compare_cn_track.Rd	Points documentation source to new plotting file.
man/compare_cn_track_summary.Rd	New Rd for `compare_cn_track_summary`.
man/call_hmm_dosages.Rd	Updates documented defaults for het gating.
docs/QuickStart_template.Rmd	New/expanded QuickStart covering new metrics and utilities.
docs/Qploidy_alfalfa_tutorial.Rmd	Tutorial updates to reflect new workflow/tools and defaults.
DESCRIPTION	Bumps package version to 1.18.3.

Files not reviewed (10)

man/call_hmm_dosages.Rd: Generated file
man/compare_cn_track.Rd: Generated file
man/compare_cn_track_summary.Rd: Generated file
man/compute_baf_likelihoods.Rd: Generated file
man/count_types.Rd: Generated file
man/em_hmm_cn.Rd: Generated file
man/filter_hmm_CN.Rd: Generated file
man/hmm_estimate_CN.Rd: Generated file
man/plot_karyotype.Rd: Generated file
man/print.count_types.Rd: Generated file

Suppressed comments (1)

R/plots_hmm_results_all_samples.R:25

The roxygen lines for facet_nrow and gray_CN use ##' instead of #', so these parameters likely won’t be included in the generated documentation.

#' @param facet_ncol Number of columns for facet_wrap. If NULL, will be determined automatically unless facet_nrow is set.
##' @param facet_nrow Number of rows for facet_wrap. If set, facet_ncol will be determined automatically unless both are set.
##' @param gray_CN Integer or NULL. If provided, this copy-number value is colored gray (used as the baseline color). All CNs below it are colored blue and above red. If NULL (default), the baseline is auto-detected as the most frequent CN weighted by window length.
#' @param add_het Logical. If TRUE, a heterozygosity column is added to the right side of the plot, with one square per sample colored from blue (low) to red (high). Requires \code{hmm_dosage_calls}. Default is TRUE.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

+
+# Qploidy2 1.18.2
+
+* Revert mu stability guard introduced in 1.18.0: `em_hmm_cn` now always updates mu unconditionally (standard EM update with `pmax` denominator guard). The while-loop in `hmm_estimate_CN` already handles non-monotonic mu.


+    het_raw <- hmm_dosage_calls |>
+      filter(.data$SampleName %in% samples_present)
+    if (!is.null(chromosomes)) {
+      het_raw <- het_raw |> filter(.data$Chr %in% chromosomes)
+    }
+    het_df <- het_raw |>


Copilot

Pull request overview

Copilot reviewed 23 out of 34 changed files in this pull request and generated 1 comment.

Files not reviewed (10)

man/call_hmm_dosages.Rd: Generated file
man/compare_cn_track.Rd: Generated file
man/compare_cn_track_summary.Rd: Generated file
man/compute_baf_likelihoods.Rd: Generated file
man/count_types.Rd: Generated file
man/em_hmm_cn.Rd: Generated file
man/filter_hmm_CN.Rd: Generated file
man/hmm_estimate_CN.Rd: Generated file
man/plot_karyotype.Rd: Generated file
man/print.count_types.Rd: Generated file

Suppressed comments (3)

Previously missed (3) — in code that hasn't changed since the last review.

R/hmm_main.R:817

w_baf_use / baf_weight will produce NaN (0/0) when baf_weight is set to 0 (which is allowed by the documented 0–1 range), causing CN_reliability to become NA for all windows. Add an explicit guard for non-finite/zero baf_weight before dividing.

  w_baf_use <- if (z_only) rep(0, W) else w_baf
  w_baf_norm <- pmax(z_no_baf_scale, pmin(1, w_baf_use / baf_weight))
  CN_reliability <- post_max * w_baf_norm

R/standardization.R:809

mode(hmm_CN_multi$by_marker$CN_call) can return NA when CN_call contains many missing values (e.g., after filter_hmm_CN masking). In that case, ploidy.standardization != mode(...) becomes NA and the else if will error with “missing value where TRUE/FALSE needed”, breaking re_standardize() for filtered inputs. Consider using an NA-safe mode (ignore NAs) and explicitly error if all CN calls are NA before doing the comparison/override.

    if (is.null(ploidy.standardization)) {
      ploidy.standardization <- mode(hmm_CN_multi$by_marker$CN_call)
      if (verbose) cat("ploidy.standardization not provided, using:", ploidy.standardization, "\n")
    } else if (ploidy.standardization != mode(hmm_CN_multi$by_marker$CN_call)) {
      warning(sprintf(

R/filter_hmm_CN.R:72

When a user supplies a threshold but the corresponding column is missing in hmm_CN$by_window (e.g., older objects without CN_reliability), the filter is silently skipped and the function can still print “No windows failed…”, which is misleading. Emit a warning when a requested filter can’t be applied due to missing columns.

  if (!is.null(min_CN_reliability) && "CN_reliability" %in% names(bw))
    fail <- fail | is.na(bw$CN_reliability) | bw$CN_reliability < min_CN_reliability

  if (!is.null(min_post_max) && "post_max" %in% names(bw))
    fail <- fail | is.na(bw$post_max) | bw$post_max < min_post_max

+    # Build per-segment tooltip lookup for het
+    het_lookup <- if (!is.null(het_df)) {
+      setNames(round(het_df$het, 3), as.character(het_df$SampleName))
+    } else NULL
+


Cristianetaniguti added 10 commits August 12, 2026 10:21

turning HMM more conservative for segments

2a6c9cb

new summary functions

0a50531

SAT-209

aefe065

update news

40aaa94

SAT-209 update documentation

686da48

up version

71e2375

v1.18.2

f3bc44d

force stand.ploidy if dosage estimated

cb09a45

put back standardization

8ee1dbf

change notations

f4f7e0d

Cristianetaniguti requested a lite review from Copilot August 21, 2026 19:28

Copilot started reviewing on behalf of Cristianetaniguti August 21, 2026 19:28 View session

Cristianetaniguti added the enhancement New feature or request label Aug 21, 2026

Copilot AI reviewed Aug 21, 2026

View reviewed changes

copilot fixes

0e362df

Cristianetaniguti requested a lite review from Copilot August 24, 2026 17:39

Copilot started reviewing on behalf of Cristianetaniguti August 24, 2026 17:39 View session

Copilot AI reviewed Aug 24, 2026

View reviewed changes

Comment thread R/plots_hmm_results_all_samples.R

Comment on lines +222 to +226

# Build per-segment tooltip lookup for het

het_lookup <- if (!is.null(het_df)) {

setNames(round(het_df$het, 3), as.character(het_df$SampleName))

} else NULL

Cristianetaniguti added 2 commits August 25, 2026 09:07

correct news

084a1bf

function for allo

b33035a

Navigation Menu

Uh oh!

Sat 206 indenpendent chr estimations - #17

Sat 206 indenpendent chr estimations#17
Cristianetaniguti wants to merge 13 commits into
developmentfrom
SAT-206-Indenpendent_chr_estimations

Cristianetaniguti commented Aug 21, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants


		# Qploidy2 1.18.2

		* Revert mu stability guard introduced in 1.18.0: `em_hmm_cn` now always updates mu unconditionally (standard EM update with `pmax` denominator guard). The while-loop in `hmm_estimate_CN` already handles non-monotonic mu.

Uh oh!

Conversation

Cristianetaniguti commented Aug 21, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants