biodosetools 3.5.0.9000

• Major rewrite of plot_estimated_dose_curve() function.
• Delta method calculations are now all done via msm::deltamethod().
• Added progress and alert notifications to UI.
• Added {markdown} as suggested dependency.
• Added {cli} as dependency.
• Tweaked references section in {pkgdown} site to only include exported functions organised by topic/usage.
• Use consistently genomic conversion factor or genome_factor across all functions, help dialogues, and reports.
• Added R-CMD check GitHub Action.
• Added vignettes.

New features

• Return conf_int in estimate_*() family of functions.
• Parse est_full_doses$type in plot_estimated_dose_curve() function so unused assessments are not shown in legend.
• Added match_names() function to match est_doses list names when calling plot_estimated_dose_curve().
• Updated plot_estimated_dose_curve() to automatically build est_full_doses object from list of dose estimation results, added parse_conf_int_text() auxiliary function.

Bug fixes

• Correct dispersion index value on estimate_whole_body_delta() when there’s no aberrations.
• Fixed issue in missing count data in DOCX fitting report.
• Removed double calculation of cov_est in estimate_partial_dolphin().
• Wrapped mixed Poisson model dose estimation in try() to ensure convergence (up to 5 tries).
• Mixed Poisson model estimate_hetero_mixed_poisson() uses exact delta methods for dose and fraction irradiated, as described in paper by Pujol et al. (2016) https://doi.org/10.1667/RR14145.1.
• Fixed gamma and gamma_error parsing when calling estimate_hetero_mixed_poisson() in mod_estimation_results_server().
• Wrap unicode characters (\uxxxx) in rlang::as_utf8_character() to avoid “unable to translate to native encoding” warning on Windows.

New functions

• get_deltamethod_std_err(): auxiliary function to wrap all msm::deltamethod() calls.

Improvements

• Fixed order of output case data columns in calculate_aberr_table() so they match the final output in Shiny.
• Renamed count-data-IAEA.csv to count-data-barquinero-1995.csv from Barquinero et al. (1995) https://doi.org/10.1016/0027-5107(94)00150-4.
• Added count-data-rodriguez-2004.csv for total translocations from Rodriguez et al. (2004) https://doi.org/10.1667/RR3198.
• Updated translocations fitting unit tests to use data from Rodriguez et al. (2004) https://doi.org/10.1667/RR3198.
• Rewrote calculate_aberr_power() to remove {purrr} dependency and reduce computation time by 4.
• Replaced message() calls with appropriate cli::cli_*() calls.
• Split quasi-Poisson from automatic fitting in fit_glm_method(), as it is otherwise confusing.
• Updated \sigma to \hat{\sigma} in fix_count_data_names() and respective tests.
• Updated unit tests to test individual values both in dose-effect fitting and dose estimation.
• Use *-fitting-results-*.rds instead of *-fitting-data-*.rds when exporting RDS objects in mod_fitting_results_server().

UI Improvements

• Added new side-widget-* CSS classes to better style file download/format buttons.
• Added widget_sep_vert() function.
• Use new sep-widget_* CSS classes in fitting UI modules.
• Updated report help modals text.
• Using new side-widget-* CSS classes in dose estimation UI modules.
• Added progress notifications to calculations via shiny::Progress.
• Updated _pkgdown.yml and simplified extra.css thanks to Bootstrap 5 support in {pkgdown} 2.0.0.

biodosetools 3.5.0 (2021-05-26)

• Added unit tests for to check code coverage and to validate that the code performs as expected.
• Major revision of the fitting and dose estimation pipelines, introducing new auxiliary functions to reduce duplicated code.
• Major simplification of report rendering process; abandoned HTML output.
• Added custom {pkgdown} theme for more consistent branding.
• Added preliminary citation to README.
• Moved packages required only for reports to Suggests field in DESCRIPTION file.

New features

• Added irradiation conditions input to dose-effect fitting modules.
• Reports include conditional formatting of u-value and other formatting refinements.
• Added biodosetools_version element to Shiny App’s exported *-fitting-data-YYY-MM-DD.rds files.
• fit() function allows to optionally select "glm" or "maxlik" algorithms. If "glm" if selected, the original tryCatch() routine will be executed.

Bug fixes

• Fixed unexported calculate_aberr_var() function.
• Stopped using weights in fitting algorithms (glm and glm.nb). Fixes #20, addresses part of #14 as well.
• Fixed aggregated count data column parsing in fix_count_data_names() function.
• Fixed case data parsing issue when .csv file is not perfectly formatted.
• Fixed issue with manual fitting curve input (fixes #23).
• Fixed bug in calculate_yield_infimum() where infima were being calculated for yield estimate only (fixes #26).
• Fixed wrong calculation of aberrations X for Ck when k>=10 (fixes #27).
• Fixed fix_count_data_names() to properly correct Ck>=C10 (related to issue #27).
• Added “where” to global variables (see https://github.com/r-lib/tidyselect/issues/201).

New functions

• inner_column(), to fix padding of boxes inside columns.
• widget_sep(), to insert div(class = "widget-sep", br()) calls.
• widget_sep_vert() to insert div(style = "height: 8px;") calls.
• names_from_model_formula(), to parse rhandsontable() Unicode row and column coefficient names.
• parse_model_formula() to get raw and TeX formulae from model_formula.
• calculate_trans_rate_sigurdson() and calculate_trans_rate_manual() to calculate translocation rates.
• calculate_aberr_table() wrapper, to calculate aberration tables for count and cases data.
• init_aberr_table(), to initialise aberration distribution tables in fitting and estimation server modules.
• list_fitting_formulas(), to replace global global_fitting_formulas object.
• generalise_fit_coeffs() and generalise_fit_var_cov_mat(), used internally in estimation functions instead of using general_fit_* as parameters.

Function refactoring

• Refactored get_decision_threshold() function to calculate_decision_threshold(), and added calculate_decision_threshold_table() wrapper function.
• Refactored calculate_decision_threshold_*() functions to remove input argument.
• Refactored get_*() translocation functions to calculate_*() for better consistency.
• Refactored get_*_dose_curve() functions to plot_*_dose_curve() for more clarity.
• Refactored get_model_statistics() to calculate_model_stats().
• Refactored get_fit_*() functions to fit() and fit_*_method().

Improvements

• Added Oliveira citation on get_fit_maxlik_method() function.
• Fixed Gaëtan’s name order in citation and contributors list.
• Updated README and About body text.
• Renamed source R files for better naming consistency.
• Changed structure of modules to mod_<calc_type>_<aberration>_*().
• Generalised get_model_statistics() function so that local get_model_statistics() definition could be removed from mod_estimate_fit_curve_server() module.
• Multiple (18 files) <aberration>-<module>-<format>.Rmd report templates have been merged into <module>-<format>.Rmd (4 files).
• Use correct_negative_vals() to ensure correct dose estimation when X < Xc in translocations assay.
• Replaced all *_at() and *_if() occurrences by their {dplyr} 1.0.0 equivalents.

UI Improvements

• Replaced column(width = X) calls by col_X() in UI modules.
• Added tabitem-container class to tabItem() page containers for fixed max-width while keeping responsive UI.
• Added col-inner-textinput-* CSS classes for Irradiation conditions textInput() widgets’ containers.
• Added sep-widget-download and sep-widget-format CSS classes to unify download and format select buttons into a single widget.

biodosetools 3.4.0 (2020-10-11)

• Removed all non-ASCII characters.
• Major overhaul of reports.
• Added “Report bug” button on home screen and cleaned up dashboard_*() code.
• Removed dependencies on {stringr} package.
• Initial utils::citation() support via CITATION file.
• Minimal version of R bumped to 3.5.0.
• Minimal version of {dplyr} bumped to 1.0.0.

Bug fixes

• Replaced C, α, β variables (and derivatives) by coeff_C, coeff_alpha, coeff_beta.
• Fixed error in genome_fraction parsing for translocations in estimate_partial_dolphin().
• Fixed count/case data standard error column names replacements.
• Fixed minipage issue when using more than 12 columns in {knitr} tables with column name replacements, by using format = "latex", escape = FALSE parameters in kable() call.
• Suppress YAML warning when using !expr in PDF reports. See https://github.com/rstudio/rstudio/issues/7545.

New functions

• fix_coeff_names(), to fix coefficient names in reports.
• fix_count_data_names(), to fix count/case data column names in PDF reports.
• to_title(), to replace stringr::str_to_title() using base R.

Reports improvements

• New PDF reports, which are now the default output format.
• Improved HTML reports styling.
• Added explicit M-FISH usage in translocation reports.
• Added assay name in report titles.

biodosetools 3.3.1 (2020-10-07)

Bug fixes

• Count and case data calculations are performed using calculate_aberr_*() functions. Fixes #8.
• Fixed missing {dplyr} namespace in n() function call.
• Fixed mismatched use of awesomeCheckbox() and switchInput() in confounders input.

New functions

• calculate_aberr_power(), which supersedes internal aberr_calc() function in server modules.
• calculate_aberr_mean().
• calculate_aberr_var().
• calculate_aberr_disp_index().
• calculate_aberr_u_value().

biodosetools 3.3.0 (2020-07-27)

Initial migration of the app into a {golem} R package. This means Biodose Tools is also available as a regular R package in case the user wants to build their own R scripts. This also reduced lots of code redundancies, and made the code a lot more robust.

New features

• App is now a {golem} R package.
• Migrated UI from {shinydashboard} to {bs4Dash}.
• Modals are now built with {bsplus} instead of {shinyBS}.
• Functions and internal data are documented using {roxygen2}.
• Added {pkgdown} support to build {biodosetools}’s website automatically.
• All custom theming is done through SASS and CSS instead of R code injection.

New functions

• New include_help() for help dialogues
• New help_modal_button() function to help build modal trigger buttons on boxes’ titles.
• New load_rmd_report() function to use in server downloadHandler().
• Moved get_genome_fraction() as its own exported function.

Fixes

• Fixed #7. Overestimate error of lambda when dispersion is smaller than 1.
• Fixed missing icon on sidebar.
• Fixes in collapse button paddings.
• Fixed tabBox parameters.
• Using rlang::.data$ to avoid “no visible binding for global variable” warnings.
• Fixed bsplus::bs_modal() calls in output$estimate_results_ui.
• Removed input as parameter in calculation functions.
• Fixed style of help button on box/tabBox headers.
• Updated old “biodosimetry-uab” URLs to “biodosetools-team”.

Improvements

• Stop exporting auxiliary include*() functions used on the app only.
• Moved widget_label() to widgets.R.
• Refactored inner_column() function.
• Refactored UI modules names.
• Refactored server functions and renamed R scripts.
• Updated reports.
• Using load_rmd_report() on mod_estimate_results_server() module.
• Added DNA Content Fractions of Human Chromosomes (IHGSC) data.
• Updated dna_content_fractions call on mod_trans_fraction_to_full_genome_server() module.
• Added support for tabBox on help_modal_button() widget.
• Added indentation to all module rhandsontable() calls.
• Stop exporting get_decision_threshold() function temporarily.
• “Stains color options” boxes don’t use inner_column() anymore.
• Updated NEWS.md headers so they can be parsed by {pkgdown}.

UI Improvements

• Removed experiment_select selectInput() from sidebar. Every assay is listed on the “Aberration assays” sidebarMenu().
• Loading all UI modules in dashboard_body().
• Navbar now uses rightUi and leftUi parameters, as well as skin = "light".
• Switched from shinyBS::bsButton() to shiny::actionButton() for calling modal dialogues.
• New CSS style sheet for {shinydashboard}.
• Deleted old CSS files for {bs4Dash}.
• New fonts CSS stylesheet.
• New CSS stylesheets for fixes of existing classes and definition of custom classes.
• Added hot-improved CSS class for rHandsontableOutput().
• Moved custom widgets classes from theming.R to custom.css.
• Tweaks in box, awesome-checkbox, and sidebar submenus CSS styles.
• Replaced theme_button_status() function by proper SASS to CSS compilation.
• Finished SASS box header colors.
• Using SASS modules to reconstruct biodose_style.css.
• Moved treeview-menu style to sidebar.scss.
• Added checkboxes SASS style.
• Added missing col-inner-* CSS classes.
• Added app version on sidebar footer with custom sticky-footer CSS class.
• Switched to a single stylesheet, being everything built by SASS modules.
• Added custom style to radiobuttons.

biodosetools 3.2.1 (2020-02-13)

Bug fixes

• Added required {pander} package to generate DOCX reports.

biodosetools 3.2.0 (2020-02-11)

All calculations functions previously provided in inst/app/calcs have been made proper functions on the package.

New functions

• calculate_yield() new wrapper of yield_fun(), R_factor(), and yield_error_fun().
• calculate_yield_infimum() function to calculate infima of yields given a curve.
• project_yield() merged version of the project_yield_estimate(), project_yield_lower(), and project_yield_upper() functions.

“New” functions

• get_decision_threshold().
• get_fit_dose_curve().
• get_fit_glm_method().
• get_fit_maxlik_method().
• get_fit_results().
• get_model_statistics().
• prepare_maxlik_count_data().
• AIC_from_data().
• correct_boundary().
• correct_conf_int().
• correct_negative_vals().
• correct_yield().
• get_estimated_dose_curve().
• protracted_g_function().
• R_factor().
• yield_error_fun().
• yield_fun().
• estimate_hetero().
• estimate_partial_dolphin().
• estimate_whole_body_delta().
• estimate_whole_body().

biodosetools 3.1.0 (2019-10-26)

Unofficial release (wasn’t changed on DESCRIPTION file). This includes some of the changes discussed with David in Stockholm (ERPW19).

New features

• Made var-cov matrix optional on dose estimation inputs.
• Fixed calculations to make var-cov matrix optional on dose estimation inputs.

Fixes

• Added additional package dependencies to DESCRIPTION file.
• Following the ISO, renamed detection limits to decision threshold.

Changes

• Disabled decision thresholds (for now).
• Hide AIC as a relative quality statistic of the dose estimation.
• Added mean and variance to count data tables in fitting module.

biodosetools 3.0.0 (2019-10-12)

The app is now available as an R package on GitHub (not submitted to CRAN yet).

New functions

• runApp() for launching Biodose Tools.
• %>% imported from {magrittr}.

Fixes

• Fixed variance calculation for count data.
• Provide fallback method for NB2 when using constraint-maxlik-optimization.

biodosetools 2.1.0 (2019-07-27)

New features

• Initial implementation of micronuclei, adding support in UI and server functions.
• Implemented negative binomial fitting calculation for micronuclei count data.
• Initial detection limit implementation for translocations.
• Added DOCX support for dicentrics and translocations.

New functions

• New bs4MyTabPanel() widget to remove unnecessary padding on tabCards’ panels.

Fixes

• Fixed colwidths for detection limits.
• Fixed unnecessary recalculation of num_cols for chromosome tables in translocations modules.
• New subdirectory structure for help dialogues.
• Added buttons and example tables in help dialogues.
• Updated documentation link.

Improvements

• Added button_fit dependency for detection limits calculation.
• Updated fitting reports (new hot structures & dynamic widths).
• Moved calculation functions from fitting and estimation modules to their own files in a new calcs directory.
• Translocations reports are complete now.

Changes

• Disable sourcing translations.R for now (until {shiny-i18n} is implemented).

biodosetools 2.0.0 (2019-07-17)

New features

• Added manual input of translocations name.
• Implemented dynamic calculation of hot width for counts, cases, and chromosome data tables.
• Implemented manual input of translocation frequency to be subtracted from the observed yield.
• Started detection limit calculation in fitting module.
• Implemented Delta method for whole-body estimation.
• Added calculation of standard error of Fg for translocations’ dose estimation module.

Fixes

• Fixed dependency of the “Calculate fitting” button in the translocation module.
• Fixed calculations for whole and partial body estimations for translocations module.
• Fixed colors on chromosome hot table example.
• Fix genomic conversion factor not being read when doing fitting.
• Fix Fp -> y when using dicentrics in dose estimation module.
• Fixed detection limit calculation.
• Fixed dose estimation in detection limits function.
• Corrected correction of FISH to Full genome when inputting manual fitting curve in dose estimation modules.
• Corrected reading of translocations fit curve values in dose estimation module.

Improvements

• Tweaked variable names and order in cases data for translocations dose estimation.
• Updated help_colors_dialog for translocations modules.
• New module for auxiliary rhandsontable() tables used in help dialogues.
• New help dialogues for translocations modules.
• Consistent use of genome_fraction in variables related to the genomic conversion factor/fraction of genome hybridised.
• ggplot2 curve generation is now wrapped inside a function on the dose estimation modules.
• Finished confounders help dialogue.
• New aberration calculation method using purrr::map_df() instead of a nested loops, which is about 10 times faster.
• Make clear in translocations dose estimation module that the coefficients are from the full genome.
• Added highly protracted exposure.
• Implemented calculation of detection limits from data.
• Added dose calculation to detection limits section.
• Changed σ to \sigma in help dialogues.
• Added units to result tables.
• Code clean-up on delta method for whole-body dose estimation.
• Updated plot CI text generation on legend to account for new Delta method and for separate whole-body & heterogeneous error method selection.
• CONTRIBUTORS.md is now displayed as a modal.
• Using ORCID for authors when possible.
• Updated institution links in contributors files.

UI Improvements

• Improved side-widget-tall bottom margins.
• Moved translocation frequency selectInput() next to Calculate Fitting button for better UX.
• New hot-improved CSS class for better hot tables’ formatting.
• All hot tables are using hot-improved class now.
• New contributors list style and custom CSS style for tables.
• Added experiment/aberration name on tab body title to avoid confusions.
• Improved “About” tab paddings.
• Improve widths of color columns in chromosome tables.
• Narrower color columns in chromosome tables.
• All hot tables’ widths are dynamic or exact now.
• Improved small action buttons style on help dialogues.
• New results color to differentiate from main color.
• Updated UI to adapt for new Delta method for whole-body estimation.
• Updated translocations UI to adapt for new Delta method for whole-body estimation.
• Added CI for Dolphin dose estimation UI.
• Added (work in progress) Micronuclei into UI.
• Hide language selector until shiny-i18n is implemented.

biodosetools 1.2.0 (2019-06-25)

New features

• Start translocations dose estimation module and report.
• Allow using M-Fish color scheme in translocations fitting module.
• Added method to convert from translocations measured by FISH to full genome.
• Implemented confounders modifications for translocations in generalEstimateCaseHotTable() module.

Improvements

• New chromosome table structure. transChromosomeTable() and transFractionToFullGenomeCalc() have been moved to transGeneralModule.R, as they are reused in the dose estimation module.
• Using new chromosome table UI and server modules for translocations dose estimation.
• Merged “repeated” fitting modules into a generalFittingModules.R file.
• Merged *FittingResults() modules into generalFittingResults() module.
• Updated translocations fitting report to new chromosome table structure.
• New generalEstimateFittingCurveHotTables() and generalEstimateFittingCurve() modules.
• New globalVariables.R with global_fitting_formulas list.
• Added translocations-specific code in generalEstimateFittingCurve() module to take into account used translocation frequency.
• New inputs for used translocation frequency and genomic conversion factor in manual fitting input in dose estimation module.
• Moved dose estimation to generalEstimateResults() module.
• Modify translocations (X - Xt) for whole-body and partial-body dose estimation.
• Added readOnly status to Xt, yt, yt_err variables on rhandsontable() for translocations case data.

UI improvements

• New innerColumn to use inside cards/boxes.
• New widgetLabel() function for labels above widgets without default label.
• Replaced old labels by new widgetLabel() function.
• Finished confounders UI in translocations dose estimation module.
• New mySwichInput() function based on shinyWidgets::switchInput() with sideLabel and custom status colors.
• Moved from awesomeCheckbox() to mySwitchInput() in all UI modules.
• Moved fitting curve input in translocations’ dose estimation module to the top.

Fixes

• Fix minor error in calculation of infimum yield values.
• Fixed “Color options” card and color_list object for chromosome table generation.
• Fixed missing type argument in get_model_statistics() in dose estimation module.
• Deleted heterogeneous estimation for translocations.

biodosetools 1.1.0 (2019-06-15)

New features

• Conditional tabs in sidebar by implementing bs4MyConditionalSidebarMenuItem() function.
• Implemented AIC calculation for different dose estimation assessments.
• Added ZIP archive with sample data for testing.
• New dynamic bs4TabCard() using renderUI() and session$ns() to show different tabs depending on user input.
• Implemented manual calculation of case parameters (N, X, y, y_est, DI, u) in dose estimation module.

Improvements

• Disabled automatic calculation of N, X, DI, u to avoid broken input issues when running the deployed app.
• Added CONTRIBUTORS.md file.
• Removed {fontawesome} package requirement.
• Added translocation frequency and genomic conversion factor to translocations fitting report.
• Chromosome table included in reports.
• Major rewrite of {ggplot2} plot in dose estimation module.
• Added Dolphin CI info into plot when using partial-body assessment.
• Changed display of CI in plot for dose estimation.
• Genomic conversion factor is used in transFittingResults() module for calculations instead of modifying the value of N in transFittingHotTable().
• Added theoretical calculation of model-specific statistics (glm method) on fitting module.
• Added renormalisation of data for proper calculation of model-specific statistics in fitting module.
• Added theoretical model-specific statistics calculation for constraint maxlik optimization method in fitting module.
• Added global get_model_statistics() to translocations fitting module.

UI improvements

• Cleaner version of topButton() for help button and help modals on fitting and dose estimation modules..
• More informative model summary dialogue in fitting module.
• Better output of genomic conversion factor.
• “Partial” is now “Partial-body”. Better titles in dose estimation results card.
• New bs4MyTabCard() function with topButton and noPadding capabilities.

Fixes

• Fixed problem with fit_formula_raw when using linear model. Disabled *-no-int models until Dolphin can deal with them.
• Fixed rendering problem in Firefox (and potentially other browsers) with count data rhandsontable() using 100% of the card’s height.
• Fixed dose estimation rhandsontables() divs on Firefox having height = 100% of the card’s own height.
• Fixed height error in cases data table in dose estimation module in Firefox.
• Fixed glm_results object issue in get_model_statistics() function.

biodosetools 1.0.0 (2019-05-24)

First public beta for laboratories and research centres.

New features

• Initial support of reports.
• Implemented selection of CI for dose estimation (83%-83% & 95%-95%).
• Added method selection for partial body exposure estimation.
• Fixed formatting of variance covariance matrices when rendered by rHandsontableOutput().
• Linear models compatibility for partial-body (Dolphin) and heterogeneous dose estimation.
• Experimental version of new Maximum Likelihood Estimation method for dose-effect fitting.
• Manual input of fitting coefficients and var/cov matrix for dose estimation.

Improvements

• Added format/MIME type validation for file inputs.
• Render fitting formula using MathJax.
• Format curve depending on selected assessment/method (use appropriate CI) on dose estimation.
• Check if yield projection into dose is mathematically possible. Negative values of yield or dose are changed to zero as well.
• Added modification to Merkle’s yield error function when using protracted exposures.
• Implemented protracted exposure for partial body dose estimation.
• Added automatic Poisson/quasi-Poisson selection on fitting.
• Correct p-values depending on fitting model dispersion.
• Added automatic correction of confidence intervals to use “simplified” Merkle’s method if necessary.

Fixes

• Fixed calculation of full genome.
• Fixed alignment of conditional error/method inputs for dose estimation.
• Changed “base” to “estimate” on CI tables for dose estimation.
• Fixed bs4TabCard() bug.

biodosetools 0.2.0 (2019-04-03)

Version presented in second Team Meeting at Munich.

New features

• Migrated UI from {shinydashboard} to {bs4Dash}.
• Added data, plot, and results export options.
• Added modal dialogues using {shinyBS}.
• Dose estimation modules for dicentrics.
• Complete implementation of dicentrics analysis.
• Experimental fitting for translocations assay.

biodosetools 0.1.0 (2018-11-29)

First draft and proof of concept presented to the RENEB team at Barcelona.

Features

• UI built using {shinydashboard}.
• Fitting modules for dicentrics.
• Dynamic input tables powered by {rhandsontable}.