geney 1.2.24__py2.py3-none-any.whl → 1.2.25__py2.py3-none-any.whl

geney/seqmat_utils.py

@@ -140,6 +140,14 @@ class SeqMat:
         end_pos = np.where(self.seqmat[self.ROW_INDS] == end)[0][0] + 1
         return self.seqmat[:, start_pos:end_pos]
 
+    def subseq_suffix(self, start):
+        start_pos = np.where(self.seqmat[self.ROW_INDS] == start)[0][0]
+        return self.seqmat[:, start_pos:]
+
+    def subseq_prefix(self, end):
+        start_pos = np.where(self.seqmat[self.ROW_INDS] == end)[0][0]
+        return self.seqmat[:, :end]
+
     def inspect(self, pos, context=500):
         condition = np.where(self.seqmat[1, :] == pos)[0][0]
         return SeqMat().set_seqmat(self.seqmat[:, max(0, condition - context):min(self.seqmat.shape[-1], condition + context + 1)])
@@ -147,6 +155,29 @@ class SeqMat:
     def rel_pos(self, pos):
         return np.where(self.seqmat[1, :] == pos)[0][0]
 
+    def orf_seqmat(self, tis_index):
+        if tis_index not in self.seqmat[1, :]:
+            return SeqMat('ATG')
+
+        temp = SeqMat().set_seqmat(self.subseq_suffix(tis_index))
+        # Ensure the sequence length is divisible by 3
+        seq_length = len(temp.seq)
+
+        if seq_length % 3 != 0:
+            temp.seqmat = temp.seqmat[:, :-(seq_length % 3)]  # Trim the extra nucleotides
+
+        if temp.seq[:3] == 'ATG':
+            for i in range(3, len(temp.seq), 3):
+                codon = temp.seq[i:i + 3]
+                if codon in ['TAA', 'TAG', 'TGA']:
+                    index = temp.seqmat[1, i-3]
+                    return SeqMat().set_seqmat(temp.subseq_prefix(index))  # Not include the stop codon
+
+            # If no stop codon is found, return the full sequence
+            return SeqMat().set_seqmat(temp.seq)
+
+        else:
+            return SeqMat('ATG')
 
 class Gene:
     def __init__(self, gene_name='KRAS', variation=None, organism='hg38'):
@@ -371,22 +402,20 @@ class Transcript:
 
         return mature_mrna
 
-    # def find_end_codon(self):
-    #     first_stop_index = next((i for i in range(0, len(orf) - 2, 3) if orf[i:i + 3] in {"TAG", "TAA", "TGA"}),
-    #                             len(orf) - 3)
-    #     while first_stop_index % 3 != 0:
-    #         first_stop_index -= 1
-    #
-    #     orf = orf[:first_stop_index + 3]
-    #     return None
+    def find_end_codon(self, orf):
+        first_stop_index = next((i for i in range(0, len(orf) - 2, 3) if orf[i:i + 3] in {"TAG", "TAA", "TGA"}),
+                                len(orf) - 3)
+        return first_stop_index
 
     @property
     def orf(self):
         if not (hasattr(self, 'TIS') and hasattr(self, 'TTS')):
             print("Cannot create protein without set TIS and TTS values.")
             return self
+        #   If self.TIS not in seqmat, then no orf and no protein
+        return self.mature_mrna.orf_seqmat(self.TIS)
 
-        return SeqMat().set_seqmat(self.mature_mrna.raw_subseq(self.TIS, self.TTS))
+        # return SeqMat().set_seqmat(self.mature_mrna.raw_subseq(self.TIS, self.TTS))
 
     def generate_protein(self, inplace=True, domains=None):
         protein = str(Seq(self.orf.seq).translate()).replace('*', '')

geney/spliceai_utils.py

@@ -1,11 +1,10 @@
 
 #### SpliceAI Modules
-import tensorflow as tf
 from keras.models import load_model
 from pkg_resources import resource_filename
 from spliceai.utils import one_hot_encode
 import numpy as np
-import tensorflow
+import tensorflow as tf
 
 # Check if GPU is available
 if tf.config.list_physical_devices('GPU'):

{geney-1.2.24.dist-info → geney-1.2.25.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: geney
-Version: 1.2.24
+Version: 1.2.25
 Summary: A Python package for gene expression modeling.
 Home-page: https://github.com/nicolaslynn/geney
 Author: Nicolas Lynn

{geney-1.2.24.dist-info → geney-1.2.25.dist-info}/RECORD

@@ -9,8 +9,8 @@ geney/mutation_utils.py,sha256=C-K8F8wyN5joI3ZuP-d7IMYTI43YPDXUc3IgAJ07o8Q,1546
 geney/oncosplice.py,sha256=3jJc1-CWubH2ElHEjyQtsr9JYVmfPQEpq7EX-IfY-t8,20806
 geney/pangolin_utils.py,sha256=S2uMjQnnxqWSnfuMaEjo-wq52DVKFiXt__L5VPdtzyU,2939
 geney/power_utils.py,sha256=MehZFUdkJ2EFUot709yPEDxSkXmH5XevMebX2HD768A,7330
-geney/seqmat_utils.py,sha256=4MiN6rGeQMfWK6bXOHGddxBffx8v4sT1THkZe-AceXE,15611
-geney/spliceai_utils.py,sha256=BiTRIfrovX9qo9xup6bFWp0qkvmW9NVPY98Zw8-OaL0,1891
+geney/seqmat_utils.py,sha256=JZk-dY7SjHESK1NaXfWuDpZGLJkeT_03ZfZS6gHBEr0,16796
+geney/spliceai_utils.py,sha256=gIGPC8u3J15A7EQrk2Elho5PbF9MmUUNopGGH-eEV8s,1873
 geney/splicing_utils.py,sha256=pS3jZEpmnDkbT1jjaJh-O6I--Xm22e5dj-GQu7IAZSQ,15943
 geney/survival_utils.py,sha256=2CAkC2LsspicHIdrqsiPnjgvpr5KHDUfLFFqnRbPJqs,5762
 geney/tcga_utils.py,sha256=vXSMf1OxoF_AdE_rMguy_BoYaart_E1t4FFMx2DS1Ak,15585
@@ -19,7 +19,7 @@ geney/translation_initiation/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NM
 geney/translation_initiation/tis_utils.py,sha256=iXrWVijyPe-f8I9rEVGdxNnXBrOGPoKFjmvaOEnQYNE,4446
 geney/translation_initiation/resources/kozak_pssm.json,sha256=pcd0Olziutq-6H3mFWDCD9cujQ_AlZO-iiOvBl82hqE,1165
 geney/translation_initiation/resources/tis_regressor_model.joblib,sha256=IXb4DUDhJ5rBDKcqMk9zE3ECTZZcdj7Jixz3KpoZ7OA,2592025
-geney-1.2.24.dist-info/METADATA,sha256=sJGrKawFcaFyF1QwLOHn7dNYeznt5f2fKL7NDZvqqq8,948
-geney-1.2.24.dist-info/WHEEL,sha256=fS9sRbCBHs7VFcwJLnLXN1MZRR0_TVTxvXKzOnaSFs8,110
-geney-1.2.24.dist-info/top_level.txt,sha256=O-FuNUMb5fn9dhZ-dYCgF0aZtfi1EslMstnzhc5IIVo,6
-geney-1.2.24.dist-info/RECORD,,
+geney-1.2.25.dist-info/METADATA,sha256=RUiiqgqgJ0Qr9_iJvEds3xzIrFhPiW9qd-DBkdb2SxI,948
+geney-1.2.25.dist-info/WHEEL,sha256=fS9sRbCBHs7VFcwJLnLXN1MZRR0_TVTxvXKzOnaSFs8,110
+geney-1.2.25.dist-info/top_level.txt,sha256=O-FuNUMb5fn9dhZ-dYCgF0aZtfi1EslMstnzhc5IIVo,6
+geney-1.2.25.dist-info/RECORD,,