Quantitative, Multispecies Monitoring at a Continental Scale

Read me

code/

Core analysis code.

• R code

  • 1_Run_QM_qPCR.qmd Runs the joint model that connects qPCR (absolute for the reference species) with metabarcoding (compositional) and produces model-ready objects + posterior outputs.

  • 2_sdmTMB_smooths_13sp.qmd Fits spatial (sdmTMB) smooths to the joint-model concentration estimates and produces map-ready outputs.

  • 3_All_Figures.qmd Produces final figures from model outputs (joint model + smooths).

• Stan model code (used by 1_Run_QM_qPCR.qmd):

  • Joint_model.stan — Stan program implementing the joint statistical model linking qPCR (absolute, reference species) with metabarcoding (compositional, multi-species).

  • Mock_model.stan — Stan program for the mock-community calibration component (e.g., amplification efficiency / bias parameters).

  • Mock_model.rds — saved fitted object / calibration output associated with the mock model (generated by the workflow; kept here for reuse).

• R helper scripts

  • load_QM-qPCR_data.R — loads raw inputs from data/ and standardizes formats.

  • qm_data_prep_functions.R — shared utility functions for cleaning, joins, reshaping, QC, etc.

  • smoothers.R — helper functions for the smoothing / mapping stage.

data/

Raw and lightly processed inputs used by the workflow.

• hake_qpcr/ — qPCR inputs for the reference species.

• metabarcoding/ — metabarcoding reads for environmental samples.

• metabarcoding_db/ — taxonomy / reference database products used in assignment.

• metabarcoding_mocks/ — mock community inputs/outputs used for amplification-bias calibration.

• metadata/ — station/sample metadata (locations, depths, cruise/station IDs, etc.).

Intermediate_data/

Cached outputs produced by the pipeline (to avoid re-running expensive steps).

Important: everything in Intermediate_data/ is generated from:

• 1_Run_QM_qPCR.qmd and

• 2_sdmTMB_smooths_13sp.qmd

Files:

• Joint_mod_data_input.rds — compiled data object passed into the joint model.

• Joint_mod_output.rds — joint model results / posterior summaries.

• Mock_mod_output.rds — mock-community calibration outputs (amplification efficiencies / bias correction objects).

• Log_D_est.rds — species-by-sample estimated log DNA concentrations from the joint model.

• Log_D_est_smoothed.rds — smoothed (sdmTMB) versions of Log_D_est.

• all_maps_se.rds — standard errors / uncertainty layers for mapped smooth predictions.

Recommended run order

1. Run the joint model + calibration

   • Render / run: code/Stan_models/1_Run_QM_qPCR.qmd

   • Writes model inputs/outputs to Intermediate_data/

2. Fit spatial smooths

   • Render / run: code/Stan_models/2_sdmTMB_smooths_13sp.qmd

   • Writes smoothed outputs and map uncertainty objects to Intermediate_data/

3. Build final figures

   • Render / run: code/Stan_models/3_All_Figures.qmd

   • Uses Intermediate_data/ as inputs

Notes

• This is an RStudio project: open QM-qPCR-joint.Rproj to get the correct working directory.