phykit 2.1.27__tar.gz → 2.1.30__tar.gz

{phykit-2.1.27 → phykit-2.1.30}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: phykit
-Version: 2.1.27
+Version: 2.1.30
 Home-page: https://github.com/jlsteenwyk/phykit
 Author: Jacob L. Steenwyk
 Author-email: jlsteenwyk@gmail.com

{phykit-2.1.27 → phykit-2.1.30}/phykit/cli_registry.py

@@ -88,6 +88,8 @@ ALIAS_TO_HANDLER: Dict[str, str] = {
     "pglm": "phylogenetic_glm",
     "asr": "ancestral_state_reconstruction",
     "anc_recon": "ancestral_state_reconstruction",
+    "conc_asr": "concordance_asr",
+    "casr": "concordance_asr",
     "simmap": "stochastic_character_map",
     "scm": "stochastic_character_map",
     "contmap": "cont_map",
@@ -146,6 +148,10 @@ ALIAS_TO_HANDLER: Dict[str, str] = {
     "tness": "treeness",
     # Alignment- and tree-based aliases
     "sat": "saturation",
+    "threshold": "threshold_model",
+    "thresh": "threshold_model",
+    "threshbayes": "threshold_model",
+    "thresh_bayes": "threshold_model",
     "toverr": "treeness_over_rcv",
     "tor": "treeness_over_rcv",
     # Helper aliases

{phykit-2.1.27 → phykit-2.1.30}/phykit/phykit.py

@@ -163,6 +163,9 @@ class Phykit:
                 ancestral_state_reconstruction (alias: asr; anc_recon)
                     - estimate ancestral states for continuous traits using
                       ML (fast or VCV-based) with optional contMap plot
+                concordance_asr (alias: conc_asr; casr)
+                    - concordance-aware ancestral state reconstruction
+                      incorporating gene tree discordance
                 bipartition_support_stats (alias: bss)
                     - calculates summary statistics for bipartition support
                 branch_length_multiplier (alias: blm)
@@ -240,6 +243,9 @@ class Phykit:
                       gamma statistic for diversification rate testing
                 network_signal (alias: netsig; net_signal)
                     - phylogenetic signal on a network
+                threshold_model (alias: threshold; thresh; threshbayes)
+                    - Felsenstein (2012) threshold model for
+                      trait correlation via MCMC
                 polytomy_test (alias: polyt_test; polyt; ptt)
                     - conducts a polytomy test using gene
                       support frequencies
@@ -1541,6 +1547,98 @@ class Phykit:
         _add_json_argument(parser)
         _run_service(parser, argv, AncestralReconstruction)
 
+    @staticmethod
+    def concordance_asr(argv):
+        parser = _new_parser(
+            description=textwrap.dedent(
+                f"""\
+                {help_header}
+
+                Concordance-aware ancestral state reconstruction (ASR)
+                that incorporates gene tree discordance into ancestral
+                estimates. Standard ASR operates on a single species tree
+                and ignores gene tree conflict. This command propagates
+                topological uncertainty from gene tree discordance into
+                ancestral state estimates using gene concordance factors.
+
+                Two methods are available:
+                - weighted (default): concordance-weighted ASR on the
+                  species tree with NNI alternatives, using law of total
+                  variance to separate topological vs parameter uncertainty
+                - distribution: reconstruct on each gene tree independently,
+                  map nodes by descendant-set identity, report concordance-
+                  weighted means and percentile CIs
+
+                Aliases:
+                  concordance_asr, conc_asr, casr
+                Command line interfaces:
+                  pk_concordance_asr, pk_conc_asr, pk_casr
+
+                Usage:
+                phykit concordance_asr -t <species_tree> -g <gene_trees> -d <trait_data>
+                    [-c <trait>] [-m weighted|distribution] [--ci]
+                    [--plot <output>] [--missing-taxa error|shared] [--json]
+
+                Options
+                =====================================================
+                -t/--tree                   species tree file
+
+                -g/--gene-trees             file with gene trees (multi-
+                                            Newick, one per line)
+
+                -d/--trait_data             trait data file (2-column or
+                                            multi-trait with header)
+
+                -c/--trait                  trait column name (required
+                                            for multi-trait files)
+
+                -m/--method                 method: weighted or distribution
+                                            (default: weighted)
+
+                --ci                        include 95%% confidence
+                                            intervals
+
+                --plot                      output path for concordance
+                                            ASR plot
+
+                --missing-taxa              how to handle taxa mismatches:
+                                            shared (default) or error
+
+                --json                      output results as JSON
+                """
+            ),
+        )
+        parser.add_argument(
+            "-t", "--tree", type=str, required=True, help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "-g", "--gene-trees", type=str, required=True, help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "-d", "--trait_data", type=str, required=True, help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "-c", "--trait", type=str, required=False, default=None,
+            help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "-m", "--method", type=str, required=False, default="weighted",
+            choices=["weighted", "distribution"], help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "--ci", action="store_true", required=False, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--plot", type=str, required=False, default=None,
+            help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "--missing-taxa", type=str, required=False, default="shared",
+            choices=["error", "shared"], help=SUPPRESS, metavar=""
+        )
+        _add_json_argument(parser)
+        _run_service(parser, argv, ConcordanceAsr)
+
     @staticmethod
     def bipartition_support_stats(argv):
         parser = _new_parser(
@@ -2291,6 +2389,10 @@ class Phykit:
                 -p/--permutations           number of permutations for
                                             blombergs_k (default: 1000)
 
+                -g/--gene-trees             optional multi-Newick file of
+                                            gene trees for discordance-aware
+                                            VCV computation
+
                 --json                      optional argument to output
                                             results as JSON
                 """
@@ -2321,6 +2423,10 @@ class Phykit:
             help=SUPPRESS,
             metavar="",
         )
+        parser.add_argument(
+            "-g", "--gene-trees", type=str, required=False, default=None,
+            help=SUPPRESS, metavar=""
+        )
         _add_json_argument(parser)
         _run_service(parser, argv, PhylogeneticSignal)
 
@@ -2405,6 +2511,10 @@ class Phykit:
                 --plot-output               output path for plot
                                             (default: phylo_ordination_plot.png)
 
+                -g/--gene-trees             optional multi-Newick file of
+                                            gene trees for discordance-aware
+                                            VCV computation
+
                 --json                      optional argument to output
                                             results as JSON
                 """
@@ -2509,6 +2619,10 @@ class Phykit:
             required=False,
             help=SUPPRESS,
         )
+        parser.add_argument(
+            "-g", "--gene-trees", type=str, required=False, default=None,
+            help=SUPPRESS, metavar=""
+        )
         _add_json_argument(parser)
         _run_service(parser, argv, PhylogeneticOrdination)
 
@@ -2662,6 +2776,10 @@ class Phykit:
                 -m/--method                 method to use: BM or lambda
                                             (default: BM)
 
+                -g/--gene-trees             optional multi-Newick file of
+                                            gene trees for discordance-aware
+                                            VCV computation
+
                 --json                      optional argument to output
                                             results as JSON
                 """
@@ -2695,6 +2813,10 @@ class Phykit:
             help=SUPPRESS,
             metavar="",
         )
+        parser.add_argument(
+            "-g", "--gene-trees", type=str, required=False, default=None,
+            help=SUPPRESS, metavar=""
+        )
         _add_json_argument(parser)
         _run_service(parser, argv, PhylogeneticRegression)
 
@@ -2758,6 +2880,10 @@ class Phykit:
                 --log-alpha-bound           bound on log(alpha*Tmax) for
                                             logistic model (default: 4)
 
+                -g/--gene-trees             optional multi-Newick file of
+                                            gene trees for discordance-aware
+                                            VCV computation
+
                 --json                      optional argument to output
                                             results as JSON
                 """
@@ -2815,6 +2941,10 @@ class Phykit:
             help=SUPPRESS,
             metavar="",
         )
+        parser.add_argument(
+            "-g", "--gene-trees", type=str, required=False, default=None,
+            help=SUPPRESS, metavar=""
+        )
         _add_json_argument(parser)
         _run_service(parser, argv, PhylogeneticGLM)
 
@@ -3258,6 +3388,10 @@ class Phykit:
                 --models                    comma-separated list of models
                                             to fit (default: all 7)
 
+                -g/--gene-trees             optional multi-Newick file of
+                                            gene trees for discordance-aware
+                                            VCV computation
+
                 --json                      optional argument to output
                                             results as JSON
                 """
@@ -3273,6 +3407,10 @@ class Phykit:
             "--models", type=str, required=False, default=None,
             help=SUPPRESS, metavar=""
         )
+        parser.add_argument(
+            "-g", "--gene-trees", type=str, required=False, default=None,
+            help=SUPPRESS, metavar=""
+        )
         _add_json_argument(parser)
         _run_service(parser, argv, FitContinuous)
 
@@ -4019,6 +4157,100 @@ class Phykit:
         _add_json_argument(parser)
         _run_service(parser, argv, RelativeRateTest)
 
+    @staticmethod
+    def threshold_model(argv):
+        parser = _new_parser(
+            description=textwrap.dedent(
+                f"""\
+                {help_header}
+
+                Felsenstein (2012) threshold model.
+
+                Estimates the correlation between two traits (binary
+                discrete and/or continuous) using a latent-liability
+                Brownian motion model and MCMC sampling.  Binary
+                discrete characters are modelled as arising from
+                continuous liabilities that cross a threshold at 0.
+
+                This is the equivalent of phytools::threshBayes in R.
+
+                Aliases:
+                  threshold_model, threshold, thresh, threshbayes,
+                  thresh_bayes
+                Command line interfaces:
+                  pk_threshold_model, pk_threshold, pk_thresh,
+                  pk_threshbayes, pk_thresh_bayes
+
+                Usage:
+                phykit threshold_model -t <tree> -d <trait_data>
+                    --traits <trait1,trait2> --types <type1,type2>
+                    [--ngen 100000] [--sample 100] [--burnin 0.2]
+                    [--seed <int>] [--plot <file>] [--json]
+
+                Options
+                =====================================================
+                -t/--tree                   a rooted phylogeny file
+                                            with branch lengths
+                                            (required)
+
+                -d/--trait-data             tab-delimited trait file
+                                            with header row (required)
+
+                --traits                    comma-separated pair of
+                                            trait column names
+
+                --types                     comma-separated pair of
+                                            trait types (discrete or
+                                            continuous)
+
+                --ngen                      number of MCMC generations
+                                            (default: 100000)
+
+                --sample                    sample frequency
+                                            (default: 100)
+
+                --burnin                    burn-in fraction
+                                            (default: 0.2)
+
+                --seed                      random seed for
+                                            reproducibility
+
+                --plot                      output filename for
+                                            trace plots (optional)
+
+                --json                      optional argument to output
+                                            results as JSON
+                """
+            ),
+        )
+        parser.add_argument("-t", "--tree", type=str, required=True, help=SUPPRESS)
+        parser.add_argument(
+            "-d", "--trait-data", dest="trait_data", type=str, required=True, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--traits", type=str, required=True, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--types", type=str, required=True, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--ngen", type=int, default=100000, required=False, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--sample", type=int, default=100, required=False, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--burnin", type=float, default=0.2, required=False, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--seed", type=int, default=None, required=False, help=SUPPRESS
+        )
+        parser.add_argument(
+            "--plot", dest="plot_output", type=str, default=None, required=False, help=SUPPRESS
+        )
+        _add_json_argument(parser)
+        _run_service(parser, argv, ThresholdModel)
+
     @staticmethod
     def rename_tree_tips(argv):
         parser = _new_parser(
@@ -4852,6 +5084,10 @@ def ancestral_state_reconstruction(argv=None):
     Phykit.ancestral_state_reconstruction(sys.argv[1:])
 
 
+def concordance_asr(argv=None):
+    Phykit.concordance_asr(sys.argv[1:])
+
+
 def bipartition_support_stats(argv=None):
     Phykit.bipartition_support_stats(sys.argv[1:])
 
@@ -5008,6 +5244,10 @@ def relative_rate_test(argv=None):
     Phykit.relative_rate_test(sys.argv[1:])
 
 
+def threshold_model(argv=None):
+    Phykit.threshold_model(sys.argv[1:])
+
+
 def rename_tree_tips(argv=None):
     Phykit.rename_tree_tips(sys.argv[1:])
 

{phykit-2.1.27 → phykit-2.1.30}/phykit/service_factories.py

@@ -43,6 +43,7 @@ SumOfPairsScore = _LazyServiceFactory("phykit.services.alignment.sum_of_pairs_sc
 VariableSites = _LazyServiceFactory("phykit.services.alignment.variable_sites", "VariableSites")
 
 AncestralReconstruction = _LazyServiceFactory("phykit.services.tree.ancestral_reconstruction", "AncestralReconstruction")
+ConcordanceAsr = _LazyServiceFactory("phykit.services.tree.concordance_asr", "ConcordanceAsr")
 
 # Tree service loaders
 BipartitionSupportStats = _LazyServiceFactory("phykit.services.tree.bipartition_support_stats", "BipartitionSupportStats")
@@ -79,6 +80,7 @@ QuartetNetwork = _LazyServiceFactory("phykit.services.tree.quartet_network", "Qu
 NetworkSignal = _LazyServiceFactory("phykit.services.tree.network_signal", "NetworkSignal")
 LTT = _LazyServiceFactory("phykit.services.tree.ltt", "LTT")
 RelativeRateTest = _LazyServiceFactory("phykit.services.tree.relative_rate_test", "RelativeRateTest")
+ThresholdModel = _LazyServiceFactory("phykit.services.tree.threshold_model", "ThresholdModel")
 PolytomyTest = _LazyServiceFactory("phykit.services.tree.polytomy_test", "PolytomyTest")
 PrintTree = _LazyServiceFactory("phykit.services.tree.print_tree", "PrintTree")
 PruneTree = _LazyServiceFactory("phykit.services.tree.prune_tree", "PruneTree")

{phykit-2.1.27 → phykit-2.1.30}/phykit/services/tree/__init__.py

@@ -36,7 +36,9 @@ _EXPORTS = {
     "TipToTipNodeDistance": "tip_to_tip_node_distance",
     "TotalTreeLength": "total_tree_length",
     "Treeness": "treeness",
+    "ThresholdModel": "threshold_model",
     "TreenessOverRCV": "treeness_over_rcv",
+    "ConcordanceAsr": "concordance_asr",
 }
 
 __all__ = list(_EXPORTS.keys())