phykit 2.1.70__tar.gz → 2.1.72__tar.gz

{phykit-2.1.70 → phykit-2.1.72}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: phykit
-Version: 2.1.70
+Version: 2.1.72
 Home-page: https://github.com/jlsteenwyk/phykit
 Author: Jacob L. Steenwyk
 Author-email: jlsteenwyk@gmail.com

{phykit-2.1.70 → phykit-2.1.72}/phykit/cli_registry.py

@@ -55,6 +55,8 @@ ALIAS_TO_HANDLER: Dict[str, str] = {
     "sum_of_pairs_score": "sum_of_pairs_score",
     "sops": "sum_of_pairs_score",
     "sop": "sum_of_pairs_score",
+    "phylo_gwas": "phylo_gwas",
+    "pgwas": "phylo_gwas",
     "vs": "variable_sites",
     # Tree aliases
     "bss": "bipartition_support_stats",
@@ -186,6 +188,8 @@ ALIAS_TO_HANDLER: Dict[str, str] = {
     "da": "discordance_asymmetry",
     "spec_disc": "spectral_discordance",
     "sd": "spectral_discordance",
+    "impute": "phylo_impute",
+    "phylo_imp": "phylo_impute",
     "trait_correlation": "trait_correlation",
     "trait_corr": "trait_correlation",
     "phylo_corr": "trait_correlation",

{phykit-2.1.70 → phykit-2.1.72}/phykit/phykit.py

@@ -167,6 +167,9 @@ class Phykit:
                     - calculate sum-of-pairs score between a reference and query alignment
                 thread_dna (alias: pal2nal; p2n)
                     - thread dna sequences over a protein alignment
+                phylo_gwas (alias: pgwas)
+                    - phylogenetic genome-wide association study
+                      (Pease et al. 2016 approach)
                 variable_sites (alias: vs)
                     - calculates the number and percentage of variable sites
                       in an alignment
@@ -224,6 +227,9 @@ class Phykit:
                 trait_correlation (alias: trait_corr; phylo_corr)
                     - compute phylogenetic correlations between all pairs
                       of traits and display as a heatmap
+                phylo_impute (alias: impute; phylo_imp)
+                    - impute missing trait values using phylogenetic
+                      relationships and between-trait correlations
                 phylogenetic_ordination (alias: phylo_ordination; ordination; ord;
                     phylo_pca; phyl_pca; ppca; phylo_dimreduce; dimreduce; pdr)
                     - phylogenetic ordination (PCA, t-SNE, or UMAP) on
@@ -1878,6 +1884,115 @@ class Phykit:
         _add_json_argument(parser)
         _run_service(parser, argv, VariableSites)
 
+    @staticmethod
+    def phylo_gwas(argv):
+        parser = _new_parser(
+            description=textwrap.dedent(
+                f"""\
+                {help_header}
+
+                Phylogenetic genome-wide association study following the
+                Pease et al. (2016) approach.
+
+                Performs per-site association tests between alignment
+                columns and a phenotype (categorical or continuous),
+                applies Benjamini-Hochberg FDR correction, optionally
+                classifies significant associations as monophyletic or
+                polyphyletic using a phylogenetic tree, and produces a
+                Manhattan plot.
+
+                Categorical phenotypes use Fisher's exact test (2 groups)
+                or chi-squared test (>2 groups). Continuous phenotypes use
+                point-biserial correlation. Only biallelic sites are tested.
+
+                Aliases:
+                  phylo_gwas, pgwas
+                Command line interfaces:
+                  pk_phylo_gwas, pk_pgwas
+
+                Usage:
+                phykit phylo_gwas -a <alignment> -d <phenotype> -o <output>
+                  [-t <tree>] [-p <partition>] [--alpha 0.05]
+                  [--exclude-monophyletic] [--csv <file>] [--json]
+                  [shared plot options]
+
+                Options
+                =====================================================
+                -a/--alignment              FASTA alignment file
+
+                -d/--phenotype              two-column TSV file:
+                                            taxon<tab>phenotype
+
+                -o/--output                 output Manhattan plot path
+
+                -t/--tree                   optional Newick tree for
+                                            monophyletic/polyphyletic
+                                            classification
+
+                -p/--partition              optional RAxML-style partition
+                                            file for gene annotations
+
+                --alpha                     FDR significance threshold
+                                            (default: 0.05)
+
+                --exclude-monophyletic      exclude monophyletic
+                                            associations from results
+
+                --dot-size                  scale factor for dot size
+                                            in the Manhattan plot
+                                            (default: 1.0; use 2.0
+                                            for double, 0.5 for half)
+
+                --csv                       output per-site results as CSV
+                                            to the specified file
+
+                --fig-width                 figure width in inches
+                                            (auto-scaled if omitted)
+
+                --fig-height                figure height in inches
+                                            (auto-scaled if omitted)
+
+                --dpi                       resolution in DPI
+                                            (default: 300)
+
+                --no-title                  hide the plot title
+
+                --title                     custom title text
+
+                --legend-position           legend location (e.g.,
+                                            "upper right", "none")
+
+                --ylabel-fontsize           font size for y-axis labels;
+                                            0 to hide
+
+                --xlabel-fontsize           font size for x-axis labels;
+                                            0 to hide
+
+                --title-fontsize            font size for the title
+
+                --axis-fontsize             font size for axis labels
+
+                --colors                    comma-separated colors
+                                            (hex or named)
+
+                --json                      optional argument to output
+                                            results as JSON
+                """
+            ),
+        )
+        parser.add_argument("-a", "--alignment", type=str, required=True, help=SUPPRESS, metavar="")
+        parser.add_argument("-d", "--phenotype", type=str, required=True, help=SUPPRESS, metavar="")
+        parser.add_argument("-o", "--output", type=str, required=True, help=SUPPRESS, metavar="")
+        parser.add_argument("-t", "--tree", type=str, default=None, help=SUPPRESS, metavar="")
+        parser.add_argument("-p", "--partition", type=str, default=None, help=SUPPRESS, metavar="")
+        parser.add_argument("--alpha", type=float, default=0.05, help=SUPPRESS, metavar="")
+        parser.add_argument("--exclude-monophyletic", action="store_true", default=False, help=SUPPRESS)
+        parser.add_argument("--dot-size", type=float, default=1.0, help=SUPPRESS, metavar="")
+        parser.add_argument("--csv", type=str, default=None, help=SUPPRESS, metavar="")
+        add_plot_arguments(parser)
+        _add_json_argument(parser)
+        _run_service(parser, argv, PhyloGwas)
+
     @staticmethod
     def alignment_subsample(argv):
         parser = _new_parser(
@@ -3569,6 +3684,73 @@ class Phykit:
         _add_json_argument(parser)
         _run_service(parser, argv, TraitCorrelation)
 
+    @staticmethod
+    def phylo_impute(argv):
+        parser = _new_parser(
+            description=textwrap.dedent(
+                f"""\
+                {help_header}
+
+                Phylogenetic imputation of missing trait values using
+                conditional multivariate normal distributions.
+
+                Captures both phylogenetic relationships (via the
+                tree's variance-covariance matrix) and between-trait
+                correlations to predict missing values. Reports
+                imputed values with standard errors and 95% CIs.
+
+                Missing values in the input trait file may be marked
+                as NA, na, ?, or left empty.
+
+                Input is a phylogenetic tree and a tab-delimited
+                multi-trait file with a header row:
+                taxon<tab>trait1<tab>trait2<tab>...
+
+                Aliases:
+                  phylo_impute, impute, phylo_imp
+                Command line interfaces:
+                  pk_phylo_impute, pk_impute, pk_phylo_imp
+
+                Usage:
+                phykit phylo_impute -t <tree> -d <trait_data> -o <output>
+                  [-g <gene_trees>] [--json]
+
+                Options
+                =====================================================
+                -t/--tree                   tree file (required)
+
+                -d/--trait-data             multi-trait TSV with header
+                                            row; missing values marked
+                                            as NA, ?, or empty
+                                            (required)
+
+                -o/--output                 output TSV file with
+                                            imputed values (required)
+
+                -g/--gene-trees             optional multi-Newick file
+                                            of gene trees for
+                                            discordance-aware VCV
+
+                --json                      optional argument to output
+                                            results as JSON
+                """
+            ),
+        )
+        parser.add_argument(
+            "-t", "--tree", type=str, required=True, help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "-d", "--trait-data", type=str, required=True, help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "-o", "--output", type=str, required=True, help=SUPPRESS, metavar=""
+        )
+        parser.add_argument(
+            "-g", "--gene-trees", type=str, default=None, help=SUPPRESS, metavar=""
+        )
+        _add_json_argument(parser)
+        _run_service(parser, argv, PhyloImpute)
+
     @staticmethod
     def phylogenetic_ordination(argv):
         parser = _new_parser(
@@ -8163,6 +8345,10 @@ def dstatistic(argv=None):
     Phykit.dstatistic(sys.argv[1:])
 
 
+def phylo_gwas(argv=None):
+    Phykit.phylo_gwas(sys.argv[1:])
+
+
 def dfoil(argv=None):
     Phykit.dfoil(sys.argv[1:])
 
@@ -8446,5 +8632,9 @@ def tree_space(argv=None):
     Phykit.tree_space(sys.argv[1:])
 
 
+def phylo_impute(argv=None):
+    Phykit.phylo_impute(sys.argv[1:])
+
+
 def trait_rate_map(argv=None):
     Phykit.trait_rate_map(sys.argv[1:])

{phykit-2.1.70 → phykit-2.1.72}/phykit/service_factories.py

@@ -44,6 +44,7 @@ RelativeCompositionVariability = _LazyServiceFactory("phykit.services.alignment.
 RelativeCompositionVariabilityTaxon = _LazyServiceFactory("phykit.services.alignment.rcvt", "RelativeCompositionVariabilityTaxon")
 RenameFastaEntries = _LazyServiceFactory("phykit.services.alignment.rename_fasta_entries", "RenameFastaEntries")
 SumOfPairsScore = _LazyServiceFactory("phykit.services.alignment.sum_of_pairs_score", "SumOfPairsScore")
+PhyloGwas = _LazyServiceFactory("phykit.services.alignment.phylo_gwas", "PhyloGwas")
 VariableSites = _LazyServiceFactory("phykit.services.alignment.variable_sites", "VariableSites")
 
 AncestralReconstruction = _LazyServiceFactory("phykit.services.tree.ancestral_reconstruction", "AncestralReconstruction")
@@ -111,6 +112,7 @@ TreenessOverRCV = _LazyServiceFactory("phykit.services.tree.treeness_over_rcv",
 EvoTempoMap = _LazyServiceFactory("phykit.services.tree.evo_tempo_map", "EvoTempoMap")
 DiscordanceAsymmetry = _LazyServiceFactory("phykit.services.tree.discordance_asymmetry", "DiscordanceAsymmetry")
 SpectralDiscordance = _LazyServiceFactory("phykit.services.tree.spectral_discordance", "SpectralDiscordance")
+PhyloImpute = _LazyServiceFactory("phykit.services.tree.phylo_impute", "PhyloImpute")
 TraitCorrelation = _LazyServiceFactory("phykit.services.tree.trait_correlation", "TraitCorrelation")
 TraitRateMap = _LazyServiceFactory("phykit.services.tree.trait_rate_map", "TraitRateMap")
 TreeSpace = _LazyServiceFactory("phykit.services.tree.tree_space", "TreeSpace")

{phykit-2.1.70 → phykit-2.1.72}/phykit/services/alignment/__init__.py

@@ -26,6 +26,7 @@ _EXPORTS = {
     "RelativeCompositionVariabilityTaxon": "rcvt",
     "RenameFastaEntries": "rename_fasta_entries",
     "SumOfPairsScore": "sum_of_pairs_score",
+    "PhyloGwas": "phylo_gwas",
     "VariableSites": "variable_sites",
 }