Merge branch 'master' into standardise-biallelic-diplotypes-471

Author: jonbrenas

malariagen_data/adir1.py

@@ -19,6 +19,10 @@
     "dirus": TAXON_PALETTE[0],
 }
 
+XPEHH_GWSS_CACHE_NAME = "adir1_xpehh_gwss_v1"
+IHS_GWSS_CACHE_NAME = "adir1_ihs_gwss_v1"
+ROH_HMM_CACHE_NAME = "adir1_roh_hmm_v1"
+
 
 class Adir1(AnophelesDataResource):
     """Provides access to data from Adir1.0 releases.
@@ -71,6 +75,10 @@ class Adir1(AnophelesDataResource):
 
     """
 
+    _xpehh_gwss_cache_name = XPEHH_GWSS_CACHE_NAME
+    _ihs_gwss_cache_name = IHS_GWSS_CACHE_NAME
+    _roh_hmm_cache_name = ROH_HMM_CACHE_NAME
+
     def __init__(
         self,
         url=None,

malariagen_data/af1.py

@@ -21,6 +21,10 @@
     "funestus": TAXON_PALETTE[0],
 }
 
+XPEHH_GWSS_CACHE_NAME = "af1_xpehh_gwss_v1"
+IHS_GWSS_CACHE_NAME = "af1_ihs_gwss_v1"
+ROH_HMM_CACHE_NAME = "af1_roh_hmm_v1"
+
 
 class Af1(AnophelesDataResource):
     """Provides access to data from Af1.x releases.
@@ -75,6 +79,7 @@ class Af1(AnophelesDataResource):
 
     _xpehh_gwss_cache_name = XPEHH_GWSS_CACHE_NAME
     _ihs_gwss_cache_name = IHS_GWSS_CACHE_NAME
+    _roh_hmm_cache_name = ROH_HMM_CACHE_NAME
 
     def __init__(
         self,

malariagen_data/ag3.py

@@ -95,6 +95,10 @@ def _setup_aim_palettes():
     "aim_species": "object",
 }
 
+XPEHH_GWSS_CACHE_NAME = "ag3_xpehh_gwss_v1"
+IHS_GWSS_CACHE_NAME = "ag3_ihs_gwss_v1"
+ROH_HMM_CACHE_NAME = "ag3_roh_hmm_v1"
+
 
 class Ag3(AnophelesDataResource):
     """Provides access to data from Ag3.x releases.
@@ -153,6 +157,7 @@ class Ag3(AnophelesDataResource):
 
     _xpehh_gwss_cache_name = XPEHH_GWSS_CACHE_NAME
     _ihs_gwss_cache_name = IHS_GWSS_CACHE_NAME
+    _roh_hmm_cache_name = ROH_HMM_CACHE_NAME
 
     def __init__(
         self,

malariagen_data/amin1.py

@@ -19,6 +19,10 @@
     "dirus": TAXON_PALETTE[0],
 }
 
+XPEHH_GWSS_CACHE_NAME = "amin1_xpehh_gwss_v1"
+IHS_GWSS_CACHE_NAME = "amin1_ihs_gwss_v1"
+ROH_HMM_CACHE_NAME = "amin1_roh_hmm_v1"
+
 
 class Amin1(AnophelesDataResource):
     """Provides access to data from Amin1.0 releases.
@@ -71,8 +75,9 @@ class Amin1(AnophelesDataResource):
 
     """
 
-    #    _xpehh_gwss_cache_name = XPEHH_GWSS_CACHE_NAME
-    #    _ihs_gwss_cache_name = IHS_GWSS_CACHE_NAME
+    _xpehh_gwss_cache_name = XPEHH_GWSS_CACHE_NAME
+    _ihs_gwss_cache_name = IHS_GWSS_CACHE_NAME
+    _roh_hmm_cache_name = ROH_HMM_CACHE_NAME
 
     def __init__(
         self,

malariagen_data/anoph/fst.py

@@ -363,7 +363,7 @@ def average_fst(
         region: base_params.region,
         cohort1_query: base_params.sample_query,
         cohort2_query: base_params.sample_query,
-        sample_query_options: Optional[base_params.sample_query] = None,
+        sample_query_options: Optional[base_params.sample_query_options] = None,
         sample_sets: Optional[base_params.sample_sets] = None,
         cohort_size: Optional[base_params.cohort_size] = fst_params.cohort_size_default,
         min_cohort_size: Optional[

malariagen_data/anoph/g123.py

@@ -133,6 +133,13 @@ def _g123_gwss(
             chunks=chunks,
         )
 
+        if gt.shape[0] < window_size:
+            raise ValueError(
+                f"Not enough sites ({gt.shape[0]}) for window size "
+                f"({window_size}). Please reduce the window size or "
+                f"use different site selection criteria."
+            )
+
         with self._spinner("Compute G123"):
             g123 = allel.moving_statistic(gt, statistic=_garud_g123, size=window_size)
             x = allel.moving_statistic(pos, statistic=np.mean, size=window_size)
@@ -240,6 +247,12 @@ def _g123_calibration(
 
         calibration_runs: Dict[str, np.ndarray] = dict()
         for window_size in self._progress(window_sizes, desc="Compute G123"):
+            if gt.shape[0] < window_size:
+                raise ValueError(
+                    f"Not enough sites ({gt.shape[0]}) for window size "
+                    f"({window_size}). Please reduce the window size or "
+                    f"use different site selection criteria."
+                )
             g123 = allel.moving_statistic(gt, statistic=_garud_g123, size=window_size)
             calibration_runs[str(window_size)] = g123
 

malariagen_data/veff.py

@@ -548,7 +548,21 @@ def _get_within_intron_effect(base_effect, intron):
             effect = base_effect._replace(effect="INTRONIC", impact="MODIFIER")
 
     else:
-        # TODO intronic INDELs and MNPs
-        effect = base_effect._replace(effect="TODO intronic indels and MNPs")
+        # INDELs and MNPs — use the closest edge of the variant to the splice site
+        if strand == "+":
+            dist_5prime = ref_start - (intron_start - 1)
+            dist_3prime = -(ref_stop - (intron_stop + 1))
+        else:
+            dist_5prime = (intron_stop + 1) - ref_stop
+            dist_3prime = -((intron_start - 1) - ref_start)
+
+        indel_min_dist = min(dist_5prime, dist_3prime)
+
+        if indel_min_dist <= 2:
+            effect = base_effect._replace(effect="SPLICE_CORE", impact="HIGH")
+        elif indel_min_dist <= 7:
+            effect = base_effect._replace(effect="SPLICE_REGION", impact="MODERATE")
+        else:
+            effect = base_effect._replace(effect="INTRONIC", impact="MODIFIER")
 
     return effect

tests/anoph/conftest.py

@@ -2260,7 +2260,7 @@ def init_hap_sites(self):
             path=path,
             contigs=self.contigs,
             snp_sites=self.snp_sites,
-            p_site=np.random.random(),
+            p_site=np.random.uniform(0.5, 1.0),
         )
 
     def init_haplotypes(self):