Merge branch 'master' into 367-haps-freq

Author: alimanfoo

docs/source/Af1.rst

@@ -71,6 +71,7 @@ SNP data access
     is_accessible
     biallelic_snp_calls
     biallelic_diplotypes
+    biallelic_snps_to_plink
 
 Haplotype data access
 ---------------------

docs/source/Ag3.rst

@@ -72,6 +72,7 @@ SNP data access
     is_accessible
     biallelic_snp_calls
     biallelic_diplotypes
+    biallelic_snps_to_plink
 
 Haplotype data access
 ---------------------

malariagen_data/__init__.py

@@ -4,6 +4,7 @@
 from .amin1 import Amin1
 from .anopheles import AnophelesDataResource, Region
 from .pf7 import Pf7
+from .pf8 import Pf8
 from .pv4 import Pv4
 from .util import SiteClass
 

malariagen_data/anoph/cnv_data.py

@@ -751,6 +751,8 @@ def plot_cnv_hmm_coverage(
         line_kwargs: Optional[gplt_params.line_kwargs] = None,
         show: gplt_params.show = True,
         output_backend: gplt_params.output_backend = gplt_params.output_backend_default,
+        gene_labels: Optional[gplt_params.gene_labels] = None,
+        gene_labelset: Optional[gplt_params.gene_labelset] = None,
     ) -> gplt_params.figure:
         debug = self._log.debug
 
@@ -782,6 +784,8 @@ def plot_cnv_hmm_coverage(
             x_range=fig1.x_range,
             show=False,
             output_backend=output_backend,
+            gene_labels=gene_labels,
+            gene_labelset=gene_labelset,
         )
 
         debug("combine plots into a single figure")
@@ -821,7 +825,6 @@ def plot_cnv_hmm_heatmap_track(
         debug = self._log.debug
 
         import bokeh.models as bkmod
-        import bokeh.palettes as bkpal
         import bokeh.plotting as bkplt
 
         region_prepped: Region = parse_single_region(self, region)
@@ -888,7 +891,7 @@ def plot_cnv_hmm_heatmap_track(
         )
 
         debug("set up palette and color mapping")
-        palette = ("#cccccc",) + bkpal.PuOr5
+        palette = cnv_params.colorscale_default
         color_mapper = bkmod.LinearColorMapper(low=-1.5, high=4.5, palette=palette)
 
         debug("plot the HMM copy number data as an image")
@@ -960,6 +963,8 @@ def plot_cnv_hmm_heatmap(
         track_height: Optional[gplt_params.track_height] = None,
         genes_height: gplt_params.genes_height = gplt_params.genes_height_default,
         show: gplt_params.show = True,
+        gene_labels: Optional[gplt_params.gene_labels] = None,
+        gene_labelset: Optional[gplt_params.gene_labelset] = None,
     ) -> gplt_params.figure:
         debug = self._log.debug
 
@@ -989,6 +994,8 @@ def plot_cnv_hmm_heatmap(
             height=genes_height,
             x_range=fig1.x_range,
             show=False,
+            gene_labels=gene_labels,
+            gene_labelset=gene_labelset,
         )
 
         debug("combine plots into a single figure")

malariagen_data/anoph/cnv_params.py

@@ -20,3 +20,6 @@
     `coverage_calls_analysis_ids` property for available values.
     """,
 ]
+
+# This is grey followed by PuOr5
+colorscale_default = ["#cccccc", "#5e3c99", "#b2abd2", "#f7f7f7", "#fdb863", "#e66101"]

malariagen_data/anoph/dipclust.py

@@ -425,16 +425,18 @@ def _dipclust_cnv_bar_trace(
             n if not pd.isna(n) else gene_ids[i] for i, n in enumerate(gene_names)
         ]
 
+        cn_mode_no_nan = np.nan_to_num(cn_mode, nan=-1)
+
         # Plot the copy number data.
         # N.B., here we have to use go.Heatmap directly rather than
         # px.imshow because the latter fails to incorporate zmin,
         # zmax and colorscale within the trace data, and so these
         # then get lost later when we try to combined into a single
         # figure.
         trace = go.Heatmap(
-            z=cn_mode,
+            z=cn_mode_no_nan,
             y=gene_labels,
-            zmin=0,
+            zmin=-1,
             zmax=4,
             colorscale=colorscale,
             showlegend=False,
@@ -564,7 +566,7 @@ def plot_diplotype_clustering_advanced(
         snp_filter_min_maf: float = 0.05,
         snp_query: Optional[base_params.snp_query] = AA_CHANGE_QUERY,
         cnv_region: Optional[base_params.regions] = None,
-        cnv_colorscale: plotly_params.color_continuous_scale = "PuOr_r",
+        cnv_colorscale: plotly_params.color_continuous_scale = cnv_params.colorscale_default,
         cnv_max_coverage_variance: cnv_params.max_coverage_variance = 0.2,
         site_mask: Optional[base_params.site_mask] = None,
         sample_sets: Optional[base_params.sample_sets] = None,

malariagen_data/anoph/distance.py

@@ -6,7 +6,6 @@
 import numba  # type: ignore
 import numpy as np
 from numpydoc_decorator import doc  # type: ignore
-import anjl.params  # type: ignore
 
 # Internal imports.
 from .snp_data import AnophelesSnpData
@@ -410,10 +409,10 @@ def plot_njt(
         metric: distance_params.distance_metric = distance_params.default_distance_metric,
         distance_sort: Optional[tree_params.distance_sort] = None,
         count_sort: Optional[tree_params.count_sort] = None,
-        center_x: anjl.params.center_x = 0,
-        center_y: anjl.params.center_y = 0,
-        arc_start: anjl.params.arc_start = 0,
-        arc_stop: anjl.params.arc_stop = 2 * math.pi,
+        center_x: distance_params.center_x = 0,
+        center_y: distance_params.center_y = 0,
+        arc_start: distance_params.arc_start = 0,
+        arc_stop: distance_params.arc_stop = 2 * math.pi,
         width: plotly_params.fig_width = 800,
         height: plotly_params.fig_height = 600,
         show: plotly_params.show = True,
@@ -426,8 +425,8 @@ def plot_njt(
         color_discrete_sequence: plotly_params.color_discrete_sequence = None,
         color_discrete_map: plotly_params.color_discrete_map = None,
         category_orders: plotly_params.category_order = None,
-        edge_legend: anjl.params.edge_legend = False,
-        leaf_legend: anjl.params.leaf_legend = True,
+        edge_legend: distance_params.edge_legend = False,
+        leaf_legend: distance_params.leaf_legend = True,
         legend_sizing: plotly_params.legend_sizing = "constant",
         thin_offset: base_params.thin_offset = 0,
         sample_sets: Optional[base_params.sample_sets] = None,
@@ -449,6 +448,10 @@ def plot_njt(
         inline_array: base_params.inline_array = base_params.inline_array_default,
         chunks: base_params.chunks = base_params.native_chunks,
     ) -> plotly_params.figure:
+        # Only import anjl if needed, as it requires a couple of seconds to compile
+        # functions.
+        import anjl  # type: ignore
+
         # Normalise params.
         if count_sort is None and distance_sort is None:
             count_sort = True

malariagen_data/anoph/distance_params.py

@@ -19,3 +19,20 @@
 ]
 
 default_nj_algorithm: nj_algorithm = "dynamic"
+
+center_x: TypeAlias = Annotated[int | float, "X coordinate where plotting is centered."]
+
+center_y: TypeAlias = Annotated[int | float, "Y coordinate where plotting is centered."]
+
+arc_start: TypeAlias = Annotated[int | float, "Angle where tree layout begins."]
+
+arc_stop: TypeAlias = Annotated[int | float, "Angle where tree layout ends."]
+
+edge_legend: TypeAlias = Annotated[
+    bool, "Show legend entries for the different edge (line) colors."
+]
+
+leaf_legend: TypeAlias = Annotated[
+    bool,
+    "Show legend entries for the different leaf node (scatter) colors and symbols.",
+]

malariagen_data/anoph/fst.py

@@ -293,6 +293,8 @@ def plot_fst_gwss(
         show: gplt_params.show = True,
         output_backend: gplt_params.output_backend = gplt_params.output_backend_default,
         clip_min: fst_params.clip_min = 0.0,
+        gene_labels: Optional[gplt_params.gene_labels] = None,
+        gene_labelset: Optional[gplt_params.gene_labelset] = None,
     ) -> gplt_params.figure:
         # gwss track
         fig1 = self.plot_fst_gwss_track(
@@ -327,6 +329,8 @@ def plot_fst_gwss(
             x_range=fig1.x_range,
             show=False,
             output_backend=output_backend,
+            gene_labels=gene_labels,
+            gene_labelset=gene_labelset,
         )
 
         # combine plots into a single figure

malariagen_data/anoph/g123.py

@@ -438,6 +438,8 @@ def plot_g123_gwss(
         output_backend: gplt_params.output_backend = gplt_params.output_backend_default,
         inline_array: base_params.inline_array = base_params.inline_array_default,
         chunks: base_params.chunks = base_params.native_chunks,
+        gene_labels: Optional[gplt_params.gene_labels] = None,
+        gene_labelset: Optional[gplt_params.gene_labelset] = None,
     ) -> gplt_params.figure:
         # gwss track
         fig1 = self.plot_g123_gwss_track(
@@ -472,6 +474,8 @@ def plot_g123_gwss(
             x_range=fig1.x_range,
             show=False,
             output_backend=output_backend,
+            gene_labels=gene_labels,
+            gene_labelset=gene_labelset,
         )
 
         # combine plots into a single figure