High-Performance Open-Source Archive
num_spline_params used in
compare_H1_and_H2() BIC calculation with the spline model
(5 parameters per curve).calculate_distance() now imputes query expression on
the registered time grid with impute_query_exp_value()
(linear interpolation via stats::approxfun()); spline-based
imputation remains available as
impute_query_exp_value_from_spline() for comparison
workflows.get_timepoint_comb_data() gains optional
cross_join_all to support four-quadrant time-point
combinations (reference-reference, query-query, etc.); ref_
/ query_ prefixes are only applied in that mode so the
default path keeps numeric time points.calculate_distance() and
plot.dist_greatR() now handle numeric and label-formatted
time-point tables more robustly (pre-aggregation relabelling removed
from distance calculation; plot-side label stripping is
conditional).dev/scripts/rebuild_brapa_registration_rds.R (plus notes
under dev/) to regenerate
inst/extdata/brapa_arabidopsis_registration.rds after
registration changes.dev/full_pipeline_example.Rmd: runnable full pipeline
(CSV → register() → summaries, curves, distances) for
smoke-testing from a source checkout.here under Suggests for the rebuild
script; updated Config/roxygen2/version for current
{roxygen2}.BIC_diff and
calculate_distance() to match current outputs.optimisation_method in
register() to be “lbfgsb” (LBFSG-B) instead of “nm”
(Nelder-Mead).arabidopsis_SOC1_data.csv and
brapa_SOC1_data.csv extdata.optimise_registration_parameters argument in
register() to use_optimisation.register() to return object of S3 class
res_greatR.calculate_distance() to return object of S3
class dist_greatR.summarise_registration() as
summary.res_greatR() S3 method.time_delta variable in registration
process.fun_args (a list of arguments used when calling
the function) in register() results.summary.res_greatR() to return NA
instead of [NA, NA] when all genes are non-registered.reg_params (table containing distribution of
registration parameters) to results list in
summary.res_greatR() method.calc_overlapping_percent() calculation.overlapping_percent when
applying manual registration.calc_variance() for data with no
replicates to consider expression_value.get_stretch_search_space_limits() and
get_shift_search_space_limits() to exclude unexplorable
regions in search space.calculate_distance() and aux
get_timepoint_comb_*_data() functions to eliminate column
selection and renaming inside lapply() calls, reducing
execution time by up to 25%.type (“registered” or “all”) and
genes_list arguments to calculate_distance()
to filter genes.plot() methods.get_shift_search_space_limits() to adjust shift
space limits accordingly to removal of time_delta variable
(see 48c943cd).overlapping_percent = 0.5 (instead of 50)
in register_manually().get_stretch_search_space_limits() to correctly
determine lower and upper limits when single stretch value is
provided.get_shift_search_space_limits() where
range variables were not available when
calc_mode == "bound".bind_results() auxiliary function to merge results from
register().theme_greatR() function and
greatR_palettes list.transform_input() S3 generic to accept different types
of input in register().plot.res_greatR() S3 method to replace
plot_registration_results().plot.dist_greatR() S3 method to replace
plot_heatmap().plot.summary.res_greatR() S3 method inspired by
WVPlots::ScatterHistC().num_cores parameter to register() to
allow users to run registration in parallel.exp_sd parameter to register() to
allow users to manually set up experimental gene expression
variance.scaling_method parameter in
register() and scale_data() to allow no
scaling (“none”, default), Z-score scaling (“z-score”), and min-max
scaling (“min-max”), and updated unit tests accordingly.register() to perform 3 sequential
registrations when using Nelder-Mead, this improves the results of
optimal stretch and shift parameters.calc_loglik() to use sigma_squared
in every time point in the sum.scaled_data() and
preprocess_data() to return all_data object
only, instead of a list() containing
all_data.compare_H1_and_H2() to return
BIC_diff column (BIC_combined - BIC_separate),
instead of BIC_combined and BIC_separate on
their own.explore_manual_search_space() to use
BIC_diff instead of BIC_combined to calculate
best_params from model_comparison table.register() to perform 3 sequential
registrations when using Nelder-Mead, this improves the results of
optimal stretch and shift parameters. This may be reverted by tweaking
neldermead() parameters to ensure correct convergence.stretch_init and shift_init
to get_search_space_limits(), and updated
optimise() to allow for different space_lims
calculation settings: automatic, given boundary box, and given initial
coords (new).mean_data calculation from
preprocess_data() and argument from
scale_data().register() to
preprocess_data() after running filter_*()
functions.results_list$data is arranged/ordered correctly
in register().get_H*_model_curves() functions to ensure model
curves are smooth.parse_gene_facets() to display
BIC_diff in facet strips.plot_mean_data parameter to
plot_registration_results().overlapping_percent parameter in
register() so it goes from 0 to 100 (it’s later normalised
in the function to avoid breakages down the line).scaling_method as an attribute in
data results from register(), this is used in
plot_registration_results() to build the y-axis label
according the the scaling method used.brapa_arabidopsis_registration.rds file with
new pipeline results.get_search_space_limits() into separate aux
functions for stretch and shift, which allows more stretch and shift
input combinations.validate_params(..., registration_type = "optimisation") to
allow more stretch and shift input combinations.get_timepoint_comb_original_data() and
get_timepoint_comb_registered_data() to perform
cross_join() on a single gene_id at a time
using lapply(), this fixes “Error: vector memory exhausted
(limit reached?)” error.match_names() to do double
setdiff() to ensure name matching is done two ways, and
updated corresponding unit test.filter_incomplete_accession_pairs() to filter out genes
that are missing one accession.calc_variance() to preprocess data variance inside
preprocess_data() instead of
calc_loglik().register_single_gene_*() functions inside
register() to simplify and generalise the pipeline for
parallel registration.calc_loglik() instead of stats::logLik().register()summarise_registration()get_approximate_stretch()plot_registration_results()plot_heatmap()calculate_distance()register() function, and
added scaling_method.register().summarise_registration(),
plot_registration_results(), plot_heatmap(),
calculate_distance() to simply require results
object from register(), vastly simplifying usage.calc_loglik_H1(),
calc_loglik_H2(), calc_overlapping_percent(),
calculate_distance(), cross_join(),
get_search_space_limits_from_params(),
get_search_space_limits(), objective_fun(),
optimise(), plot_heatmap(),
plot_registration_results(),
preprocess_data(), register_manually(),
register(), summary_registration(),
validate_params().match_names() call when validating accession
names in register()aes_string() by parsing
timepoint_var using !!ggplot2::sym()
call.preds left join in
plot_registration_results().plot_registration_results() not working
when all genes are unregistered with
type = "registered".time_delta in
preprocess_data() to ensure it’s grouped by
gene_id and accession (not just
accession).num_shifts and shift_extreme
parameters by simplified shifts parameter.calculate_between_sample_distance() to use
registration_results as primary parameter instead of
mean_df, mean_df_sc, and
imputed_mean_df.optimise_shift_extreme as
maintain_min_num_overlapping_points, properly defined and
corrected the boundary box if number overlapping points whether needed
to be maintained or not.get_approximate_stretch().x_sample
and y_sample columns according in
plot_heatmap().- character in accession names in
plot_heatmap() so that time points are parsed
correctly.optimise_registration_params().preprocess_data() to simplify
scale_and_register_data() code and reuse logic
elsewhere.get_best_stretch_and_shift_simplified().get_BIC_from_registering_data().get_boundary_box().optimise_registration_params_single_gene().optimise_registration_params() as wrapper of
optimise_registration_params_single_gene() for multiple
genes.get_best_stretch_and_shift_after_optimisation().NEWS.md file to track changes to the
package.
Need mirroring services?
Contact our team at info@vpspulse.com.
Mirror powered by VPSpulse
Infrastructure sponsored by VPSPulse & Secure Payments by ArionPay.