Package: freqTLS 0.1.0

Shinichi Nakagawa

freqTLS: Frequentist Inference for Thermal Load Sensitivity Models

A maximum-likelihood implementation of the thermal-load-sensitivity framework for thermal death-time modelling introduced by Noble, Arnold and Pottier in the 'bayesTLS' package, providing the frequentist counterpart to that Bayesian workflow. The modelling idea and the four-parameter logistic parameterisation are theirs; 'freqTLS' contributes a 'Template Model Builder' ('TMB') likelihood whose midpoint is written directly in terms of critical thermal maximum ('CTmax') and thermal sensitivity (z) so both headline quantities are estimable, and reports uncertainty through a unified trio of frequentist intervals -- Wald (delta), profile-likelihood, and bootstrap -- for binomial and beta-binomial survival counts and beta-distributed proportions. Column and formula interfaces support fixed and grouped designs plus limited independent random intercepts; prediction includes survival curves and deterministic heat-injury scenarios. Equivalence claims are restricted to the matched relative-threshold, constant-shape 'bayesTLS' configuration.

Authors:Shinichi Nakagawa [aut, cre, cph], Pieter A. Arnold [aut], Patrice Pottier [aut], Daniel W. A. Noble [aut]

freqTLS_0.1.0.tar.gz
freqTLS_0.1.0.tar.gz(r-4.7-arm64)freqTLS_0.1.0.tar.gz(r-4.7-x86_64)freqTLS_0.1.0.tar.gz(r-4.6-arm64)freqTLS_0.1.0.tar.gz(r-4.6-x86_64)
freqTLS_0.1.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
freqTLS/json (API)

# Install 'freqTLS' in R:
install.packages('freqTLS', repos = c('https://cran.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/itchyshin/freqtls/issues

Pkgdown/docs site:https://itchyshin.github.io

Uses libs:
  • c++– GNU Standard C++ Library v3
Datasets:
  • aphid_tdt - Cereal-aphid lethal-TDT data, three species across three ages
  • dsuzukii - Drosophila suzukii multi-trait thermal-tolerance data
  • shrimp_lethal - Brown shrimp lethal thermal-death-time data
  • shrimp_sublethal - Brown shrimp sublethal time-to-knockdown data
  • zebrafish_lethal - Zebrafish lethal thermal-death-time data across life stages
  • zebrafish_o2 - Zebrafish lethal-TDT data across an oxygen gradient

On CRAN:

Conda:

cpp

3.78 score 47 exports 28 dependencies

Last updated from:5b8e8dcaf3. Checks:6 OK. Indexed: yes.

TargetResultTimeFilesSyslog
linux-devel-arm64OK222
linux-devel-x86_64OK255
source / vignettesOK312
linux-release-arm64OK244
linux-release-x86_64OK293
wasm-releaseOK145

Exports:beta_binomial_tlsbeta_tlsbinomial_tlscheck_tlsclock_to_minutesderive_ctmaxderive_ltderive_tcritdiagnose_tdt_fitextract_tdtfit_4plfit_tlsformat_intervalget_ctmaxget_ctmax_drawsget_ctmax_summaryget_shapeget_tcrit_drawsget_tcrit_summaryget_zget_z_drawsget_z_summaryheat_injury_envelopemake_4pl_formulaplot_confidence_eyeplot_heat_injuryplot_survival_curvesplot_survival_surfaceplot_tdt_curvepredict_heat_injurypredict_survival_curvespredict_survival_surfaceranefsimulate_tlsstandardize_datatdt_parameter_tabletdt_quantiletidy_parameterstlstls_bftls_ctmaxtls_tcrittls_zts_cits_curvets_stage1ts_stage2

Dependencies:clicpp11farverggplot2gluegtableisobandlabelinglatticelifecyclemagrittrMASSMatrixpillarpkgconfigR6RColorBrewerRcppRcppEigenrlangS7scalestibbleTMButf8vctrsviridisLitewithr

Case study: brown-shrimp lethal TDT
The live freqTLS fit | Seeing the fit | The Confidence Eye | Survival curves | The thermal-death-time line | Deriving critical temperatures | The three-way comparison, with real numbers | Boundary: what this case study does not cover
Rendered fromcase-study-shrimp.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Case study: Drosophila suzukii lethal thermal death time across sexes
The dataset | Fitting the model: one fit, grouped by sex | The Confidence Eye | Survival curves | A lower thermal threshold: T_crit | Do the sexes differ? A frequentist contrast | Validation: does ML recover the published values? | The three-way comparison, per sex | Boundary: lethal only | Session info
Rendered fromcase-study-suzukii.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Case study: heat tolerance of three cereal aphid species (Li 2023)
The dataset and the question | The grouped fit (live) | A Confidence Eye per species | Seeing the fit | Boundary: what this case study does not cover
Rendered fromcase-study-li-aphids.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Case study: zebrafish thermal tolerance under hypoxia, normoxia, and hyperoxia
The dataset and the question | The grouped fit (live) | A Confidence Eye per treatment | Do the treatments differ? Reading the intervals | Seeing the fit | Boundary: what this case study does not cover
Rendered fromcase-study-zebrafish.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Comparing freqTLS to bayesTLS
Credit and origins | The three-way design | The reproducible recipe (not run here) | The freqTLS side (live) | The three-way comparison, with real numbers | How fast, accurate, and calibrated | Why the beta-binomial profile can dip — and what to use instead | Bootstrap fallback when a profile stays open | The teaching device: posterior density versus Confidence Eye | Beyond the matched shape: stage-specific curves | When to prefer which
Rendered fromcomparing-to-bayesTLS.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Cross-case-study summary: three taxa, one panel
The three taxa and the cross-study question | The three fits (version-stamped cache) | Assembling the cross-study table | Cross-taxon validation against bayesTLS and the two-stage estimator | The panel: a Confidence Eye per taxon and group | Within-taxon contrasts | What this shows | See bayesTLS for the Bayesian multi-taxon ridge
Rendered fromcase-study-summary.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Frequentist and Bayesian thermal-load sensitivity: strengths, weaknesses, and the blurry line
One model, two inferential philosophies | Priors: regularisation versus sensitivity | Identifiability: revealing versus masking | Non-convergence and diagnostics | Confidence versus credible intervals | Does the interval actually cover? Small-sample calibration | The blurry line: clamping, shrinkage, and penalties as implicit priors | Practical guidance: when to reach for which | References
Rendered fromfrequentist-and-bayesian.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Getting started with freqTLS
Simulate a dataset | Standardise and fit the model | Formula interface | Profile-likelihood confidence intervals | Plot the fit | Going further: groups, shape predictors, and random effects | The core workflow at a glance | Where to next
Rendered fromfreqTLS.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Heat injury under a temperature trace
What heat-injury prediction is (and is not) | Fit the curve | A temperature trace | Predict injury and survival | A repair scenario (a what-if, not an estimate) | A bootstrap envelope on survival | Absolute versus relative: what "one lethal dose" means | Extensions and boundaries
Rendered fromheat-injury.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Profile-likelihood intervals
What the profile does | Asymmetry and equivariance | Profile versus Wald | The upper asymptote up | Recovery when a profile does not close | Calibration: how well do the intervals cover? | Ship stance
Rendered fromprofile-likelihood.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Random effects: hierarchical thermal tolerance
Fixed groups vs. a random intercept | A random intercept on CTmax | Random effects on sensitivity and curve shape | Combining random effects, and what is out of scope | The likelihood view of shrinkage
Rendered fromrandom-effects.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

The model and its parameterisation
The 4PL thermal death-time curve | The direct CTmax/z parameterisation | Disjoint-bounds asymptotes | Relative versus absolute thresholds | The bridge to bayesTLS
Rendered frommodel-math.Rmdusingknitr::rmarkdown

Last update: 2026-07-21
Started: 2026-07-21

Readme and manuals

Help Manual

Help pageTopics
Cereal-aphid lethal-TDT data, three species across three agesaphid_tdt
Report identifiability diagnostics for a fitted modelcheck_tls
Convert various clock formats to minutesclock_to_minutes
Confidence intervals for a fitted thermal-load-sensitivity modelconfint.profile_tls
Derive the temperature giving a target survival at a fixed exposurederive_ctmax
Derive the lethal / survival-threshold duration at a temperaturederive_lt
Derive the critical temperature at a damage-rate floor (T_crit)derive_tcrit
Diagnose a freqTLS fit (frequentist twin of 'diagnose_tdt_fit')diagnose_tdt_fit
Drosophila suzukii multi-trait thermal-tolerance datadsuzukii
Extract z, CTmax and (optionally) T_crit with bootstrap confidence intervalsextract_tdt
Fit the 4PL thermal-load-sensitivity model by maximum likelihood (TMB)fit_4pl
Fit a single-stage 4PL thermal-load-sensitivity model by maximum likelihoodfit_tls
Format a point estimate plus confidence interval as a single stringformat_interval
Extract the CTmax estimate(s)get_ctmax
Extract the shape parameters (low, up, k, and phi)get_shape
Extract the thermal-sensitivity (z) estimate(s)get_z
Parametric-bootstrap confidence envelope for a heat-injury trajectoryheat_injury_envelope
Build a freqTLS 4PL formula from the direct CTmax/z interfacemake_4pl_formula
Confidence-Eye (or line) display of headline confidence intervalsplot_confidence_eye
Plot a heat-injury survival trajectory with a bootstrap confidence bandplot_heat_injury
Plot fitted survival curves against durationplot_survival_curves
Plot the fitted survival surface over temperature and durationplot_survival_surface
Plot the thermal death-time (TDT) curve: survival-threshold time vs temperatureplot_tdt_curve
Plot a profile-likelihood deviance curveplot.profile_tls_profile
Predict cumulative heat injury under a temperature tracepredict_heat_injury
Predict the fitted survival surface with bootstrap confidence bandspredict_survival_curves
Predict a survival surface over a temperature-by-duration gridpredict_survival_surface
Predict survival, link, or midpoint from a fitted freqTLS modelpredict.profile_tls
Profile-likelihood curves for a fitted thermal-load-sensitivity modelprint.profile_tls_profile profile.profile_tls
Random-effect BLUPs (conditional modes) for a freqTLS fitranef ranef.profile_tls
Brown shrimp lethal thermal-death-time datashrimp_lethal
Brown shrimp sublethal time-to-knockdown datashrimp_sublethal
Simulate survival-count data from the 4PL thermal-load-sensitivity modelsimulate_tls
Standardise a raw survival / proportion dataset for the TDT function librarystandardize_data
4PL parameter table (frequentist twin of 'tdt_parameter_table')tdt_parameter_table
Quantile wrapper with TDT-friendly defaultstdt_quantile
Accessors for an extract_tdt() resultget_ctmax_draws get_ctmax_summary get_tcrit_draws get_tcrit_summary get_z_draws get_z_summary tdt-accessors
Tidy the parameters of a fitted freqTLS modeltidy_parameters
Thermal-load-sensitivity quantities (z, CTmax) with confidence intervalstls tls_ctmax tls_tcrit tls_z
Build a freqTLS formula object (brms/drmTMB-style)tls_bf
Response families for thermal-load-sensitivity modelsbeta_binomial_tls beta_tls binomial_tls tls_family
Uncertainty for the classical two-stage TDT quantitiests_ci
Median LT-vs-temperature line from a two-stage fitts_curve
Stage 1 of the classical two-stage TDT pipelinets_stage1
Stage 2 of the classical two-stage TDT pipelinets_stage2
Zebrafish lethal thermal-death-time data across life stageszebrafish_lethal
Zebrafish lethal-TDT data across an oxygen gradientzebrafish_o2