hmm_estimate_CN_nQuack()now accepts CN 1 incn_grid(any subset of1:6):- CN 1 has no nQuack model. In windows whose heterozygous fraction reaches
min_het_frac, its BAF log-likelihood is set to the lowest of the other states. Otherwise BAF is neutral and z decides. - When the sample-wide heterozygous fraction is at most
min_het_fracandexp_ploidyis not given, the sample is treated as haploid (exp_ploidy = 1) and the nQuack global fit is skipped. - Windows called CN 1 get homozygous-only dosages.
- CN 0 remains unsupported on the nQuack route. It is only meaningful for array data, whose probes keep background signal without copies, and nQuack is for sequencing data only. Use
hmm_estimate_CN()for CN 0.
- CN 1 has no nQuack model. In windows whose heterozygous fraction reaches
hmm_estimate_CN(),hmm_estimate_CN_nQuack(), andhmm_estimate_CN_multi()(through...) gainmax_cn_dev.- After
exp_ploidyis set (estimated on the fullcn_grid, or given by the user),cn_gridis restricted to states withinmax_cn_devof it. For example,cn_grid = 1:6,exp_ploidy = 2,max_cn_dev = 2tests1:4. - In
hmm_estimate_CN(), CN 0 is kept whenever it is incn_grid. - This avoids false calls in mixed-ploidy populations, where a wide grid compresses the z-score spacing between states. The default,
NULL, keeps the full grid.
- After
Fixed “missing value where TRUE/FALSE needed” in
hmm_estimate_CN()andhmm_estimate_CN_nQuack()when the post-decoding override met a chromosome whose windows were allNA. It now skips such chromosomes.compare_cn_track_summary()gainssex_systemandsex_chrarguments. When not supplied, the values stored in thehmm_CNparams are used;sex_system = NULL(or"none") disables them, and other values override them. When enabled:- Samples are classified as in
count_types(), so XY males are not counted as aneuploid. - A
Sexcolumn (e.g."M (XY)","F (XX)","M (XXXY)", or"?"when unknown) is added between the matrix and the block labels.
- Samples are classified as in
Release on GitHub · About Qploidy2
Back to top