diff --git a/spine/io/parse/cluster.py b/spine/io/parse/cluster.py
index e699b16b..c583fbf8 100644
--- a/spine/io/parse/cluster.py
+++ b/spine/io/parse/cluster.py
@@ -12,7 +12,8 @@
 
 import numpy as np
 
-from spine.math.cluster import DBSCAN
+from spine.math.distance import METRICS
+from spine.math.cluster import dbscan
 
 from spine.data import Meta
 
@@ -190,8 +191,8 @@ def __init__(self, dtype, particle_event=None, add_particle_info=False,
 
         # Intialize DBSCAN if the clusters are to be broken up
         self.break_clusters = break_clusters
-        if break_clusters:
-            self.dbscan = DBSCAN(break_eps, min_samples=1, metric=break_metric)
+        self.break_eps = break_eps
+        self.break_metric_id = METRICS[break_metric]
 
         # Intialize the sparse and particle parsers
         self.sparse_parser = Sparse3DParser(dtype, sparse_event='dummy')
@@ -337,7 +338,9 @@ def process(self, cluster_event, particle_event=None,
 
                 # If requested, break cluster into detached pieces
                 if self.break_clusters:
-                    frag_labels = self.dbscan.fit_predict(voxels)
+                    frag_labels = dbscan(
+                            voxels, eps=self.break_eps, min_samples=1,
+                            metric_id=self.break_metric_id)
                     features[1] = id_offset + frag_labels
                     id_offset += max(frag_labels) + 1
 
diff --git a/spine/post/reco/direction.py b/spine/post/reco/direction.py
index f9e18fc0..d12c2307 100644
--- a/spine/post/reco/direction.py
+++ b/spine/post/reco/direction.py
@@ -91,4 +91,7 @@ def process(self, data):
 
             # Assign directions to the appropriate particles
             for i, part_id in enumerate(part_ids):
-                setattr(data[k][part_id], attrs[i], dirs[i])
+                if attrs[i].startswith('start'):
+                    setattr(data[k][part_id], attrs[i], dirs[i])
+                else:
+                    setattr(data[k][part_id], attrs[i], -dirs[i])
diff --git a/spine/utils/gnn/cluster.py b/spine/utils/gnn/cluster.py
index 5b7098b1..c95d7e15 100644
--- a/spine/utils/gnn/cluster.py
+++ b/spine/utils/gnn/cluster.py
@@ -635,7 +635,7 @@ def _get_cluster_closest_primary_label(data: nb.float64[:,:],
         # Minimize the point-cluster distances
         dists = np.empty(len(group_index), dtype=data.dtype)
         for i, c in enumerate(group_index):
-            dists[i] = np.min(sm.cdist(start_point, voxels[clusts[c]]))
+            dists[i] = np.min(sm.distance.cdist(start_point, voxels[clusts[c]]))
 
         # Label the closest cluster as the only primary cluster
         labels[group_index] = 0