Deprecate logistic_growth coloring and remove it from the workflow

The logistic growth analysis used frequencies of strains in the tree to estimate logistic growth rates. However, analyses at nextstrain.org/sars-cov-2/forecasts use all the available sequence data and provide a more sophisticated estimate of growth rate accounting for competition between circulating viruses. These updated results are available in the `mlr_lineage_fitness` analysis. The `mlr_lineage_fitness` results should be substantially more accurate.

Author: trvrb

defaults/auspice_config.json

@@ -35,16 +35,6 @@
 "title": "S1 mutations",
 "type": "continuous"
  },
- {
-"key": "logistic_growth",
-"title": "Logistic growth",
-"type": "continuous"
- },
- {
-"key": "current_frequency",
-"title": "Current frequency",
-"type": "continuous"
- },
  {
 "key": "mlr_lineage_fitness",
 "title": "MLR lineage fitness",

defaults/parameters.yaml

@@ -172,22 +172,6 @@ frequencies:
 stiffness: 20
 inertia: 0.2
 
-# Logistic growth settings
-logistic_growth:
-# Calculate logistic growth over the last N pivots which corresponds to N
-# times the amount of time represented by the `pivot_interval_units` in the
-# frequencies configuration.
-delta_pivots: 6
-
-# Set the minimum number of tips a clade must have before its logistic growth
-# is calculated.
-min_tips: 50
-
-# Set the minimum and maximum current frequency a clade must for its logistic
-# growth to be calculated.
-min_frequency: 0.000001
-max_frequency: 0.95
-
 # Cluster-specific settings
 cluster:
 # Require a minimum number of tips to be annotated as a cluster.

docs/src/reference/change_log.md

@@ -7,6 +7,8 @@ We also use this change log to document new features that maintain backward comp
 
 - 23 January 2025: Include a new `mlr_lineage_fitness` coloring. This uses live estimates of lineage fitness from [nextstrain.org/sars-cov-2/forecasts](https://nextstrain.org/sars-cov-2/forecasts) to provide a strain-level coloring to the tree. The lineage fitness estimates are automatically updated as new data comes in and so this coloring should stay current. [PR 1169](https://github.com/nextstrain/ncov/pull/1169)
 
+- 23 January 2025: Deprecate previous fitness-related colorings of `mutational_fitness`, `logistic_growth` and `current_frequency` (used by logistic growth). The new `mlr_lineage_fitess` should be more accurate / more current than these previous approaches. [PR 1169](https://github.com/nextstrain/ncov/pull/1169)
+
 ## v15 (5 December 2024)
 
 - 5 December 2024: Remove `run_pangolin` configuration option from the workflow, since the pangoLEARN tool that this enabled has been deprecated. [PR 1164](https://github.com/nextstrain/ncov/pull/1164)

docs/src/reference/workflow-config-file.rst

@@ -1097,46 +1097,6 @@ inertia
 - Unused
 
 
-logistic_growth
----------------
-
-- type: object
-- description: Parameters for estimation of logistic clade growth based on logit-transformed clade frequencies.
-- Valid attributes:
-
-.. contents::
-:local:
-
-delta_pivots
-~~~~~~~~~~~~
-
-- type: integer
-- description: Calculate logistic growth over the last N pivots which corresponds to N times the amount of time represented by the ``pivot_interval_units`` in the frequencies configuration.
-- default: ``6``
-
-min_tips
-~~~~~~~~
-
-- type: integer
-- description: The minimum number of tips a clade must have before its logistic growth is calculated.
-- default: ``50``
-
-min_frequency
-~~~~~~~~~~~~~
-
-- type: float
-- description: The minimum current frequency for a clade to have its logistic growth calculated.
-- default: ``0.000001``
-
-max_frequency
-~~~~~~~~~~~~~
-
-- type: float
-- description: The maximum current frequency for a clade to have its logistic growth calculated.
-- default: ``0.95``
-
-
-
 ancestral
 ---------
 

nextstrain_profiles/nextstrain-ci/builds.yaml

@@ -38,9 +38,3 @@ subsampling:
 frequencies:
 min_date: 2020-01-01
 max_date: 2020-05-10
-
-# Set CI-specific growth settings, such that the small tree still gets logistic
-# growth values and we properly test the growth calculations script.
-logistic_growth:
-min_tips: 0
-max_frequency: 1.0

scripts/calculate_delta_frequency.pyrenamed to scripts/deprecated/calculate_delta_frequency.py

@@ -177,16 +177,6 @@ rule auspice_config:
 "title": "ACE2 binding vs BA.2",
 "type": "continuous"
  },
- {
-"key": "logistic_growth",
-"title": "Logistic Growth",
-"type": "continuous"
- },
- {
-"key": "current_frequency",
-"title": "Current Frequency",
-"type": "continuous"
- },
  {
 "key": "mlr_lineage_fitness",
 "title": "MLR lineage fitness",

workflow/snakemake_rules/export_for_nextstrain.smk

@@ -1103,41 +1103,46 @@ rule tip_frequencies:
  --proportion-wide {params.proportion_wide} \
  --output {output.tip_frequencies_json} 2>&1 | tee {log}
  """
-
-rulelogistic_growth:
-input:
-tree="results/{build_name}/tree.nwk",
-frequencies="results/{build_name}/tip-frequencies.json",
-output:
-node_data="results/{build_name}/logistic_growth.json"
-benchmark:
-"benchmarks/logistic_growth_{build_name}.txt"
-conda:
-config["conda_environment"]
-log:
-"logs/logistic_growth_{build_name}.txt"
-params:
-method="logistic",
-attribute_name="logistic_growth",
-delta_pivots=config["logistic_growth"]["delta_pivots"],
-min_tips=config["logistic_growth"]["min_tips"],
-min_frequency=config["logistic_growth"]["min_frequency"],
-max_frequency=config["logistic_growth"]["max_frequency"],
-resources:
-mem_mb=256
-shell:
-r"""
- python3 scripts/calculate_delta_frequency.py \
- --tree {input.tree} \
- --frequencies {input.frequencies} \
- --method {params.method} \
- --delta-pivots {params.delta_pivots} \
- --min-tips {params.min_tips} \
- --min-frequency {params.min_frequency} \
- --max-frequency {params.max_frequency} \
- --attribute-name {params.attribute_name} \
- --output {output.node_data} 2>&1 | tee {log}
- """
+# This analysis is deprecated as of 2025-01-23
+# This uses frequencies of strains in the tree to estimate logistic growth rates
+# However, analyses at nextstrain.org/sars-cov-2/forecasts use all the available
+# sequence data and provide a more sophisticated estimate of growth rate accounting
+# for competition between circulating viruses. This updated results are available
+# in the `mlr_lineage_fitness` analysis
+# rule logistic_growth:
+# input:
+# tree="results/{build_name}/tree.nwk",
+# frequencies="results/{build_name}/tip-frequencies.json",
+# output:
+# node_data="results/{build_name}/logistic_growth.json"
+# benchmark:
+# "benchmarks/logistic_growth_{build_name}.txt"
+# conda:
+# config["conda_environment"]
+# log:
+# "logs/logistic_growth_{build_name}.txt"
+# params:
+# method="logistic",
+# attribute_name = "logistic_growth",
+# delta_pivots=config["logistic_growth"]["delta_pivots"],
+# min_tips=config["logistic_growth"]["min_tips"],
+# min_frequency=config["logistic_growth"]["min_frequency"],
+# max_frequency=config["logistic_growth"]["max_frequency"],
+# resources:
+# mem_mb=256
+# shell:
+# r"""
+# python3 scripts/calculate_delta_frequency.py \
+# --tree {input.tree} \
+# --frequencies {input.frequencies} \
+# --method {params.method} \
+# --delta-pivots {params.delta_pivots} \
+# --min-tips {params.min_tips} \
+# --min-frequency {params.min_frequency} \
+# --max-frequency {params.max_frequency} \
+# --attribute-name {params.attribute_name} \
+# --output {output.node_data} 2>&1 | tee {log}
+# """
 
 # This analysis is deprecated as of 2025-01-23
 # This was drawn from results at https://github.com/bkotzen/sars-cov2-modeling
@@ -1278,7 +1283,6 @@ def _get_node_data_by_wildcards(wildcards):
 rules.emerging_lineages.output.clade_data,
 rules.recency.output.node_data,
 rules.traits.output.node_data,
-rules.logistic_growth.output.node_data,
 rules.mlr_lineage_fitness.output.node_data,
 rules.distances.output.node_data,
 rules.calculate_epiweeks.output.node_data,