diff --git a/GCNFrame/Biodata.py b/GCNFrame/Biodata.py
index da14ebc..dd3f7dc 100644
--- a/GCNFrame/Biodata.py
+++ b/GCNFrame/Biodata.py
@@ -1,5 +1,7 @@
 import numpy as np
+import platform
 from Bio import SeqIO
+from concurrent.futures import ThreadPoolExecutor
 from multiprocessing import Pool
 from functools import partial
 
@@ -130,11 +132,17 @@ def __init__(self, fasta_file, label_file=None, feature_file=None, K=3, d=3, seq
     def encode(self, thread=10, save_dataset=True, save_path="./"):
         print("Encoding sequences...")
         seq_list = list(self.dna_seq.values())
-        pool = Pool(thread)
         partial_encode_seq = partial(encode_seq.matrix_encoding, K=self.K, d=self.d, seqtype=self.seqtype)
-        feature = np.array(pool.map(partial_encode_seq, seq_list))
-        pool.close()
-        pool.join()
+        
+        if platform.system() == 'Darwin':
+            with ThreadPoolExecutor(max_workers=thread) as executor:
+                feature = np.array(list(executor.map(partial_encode_seq, seq_list)))
+        else:
+            pool = Pool(thread)
+            feature = np.array(pool.map(partial_encode_seq, seq_list))
+            pool.close()
+            pool.join()
+        
         self.pnode_feature = feature.reshape(-1, self.d, 4**(self.K*2))
         self.pnode_feature = np.moveaxis(self.pnode_feature, 1, 2)
         zero_layer = feature.reshape(-1, self.d, 4**self.K, 4**self.K)[:, 0, :, :]
@@ -143,7 +151,6 @@ def encode(self, thread=10, save_dataset=True, save_path="./"):
         
         if save_dataset:
             dataset = GraphDatasetInMem(self.pnode_feature, self.fnode_feature, self.other_feature, self.edge, self.label, root=save_path)
-        
         else:
             graph = GraphDataset(self.pnode_feature, self.fnode_feature, self.other_feature, self.edge, self.label)
             dataset = graph.process()