update

TRISTAN-ORF · Nov 26, 2024 · a1adb2b · a1adb2b
1 parent 9c899b7
commit a1adb2b
Show file tree

Hide file tree

Showing 4 changed files with 8 additions and 7 deletions.
diff --git a/setup.cfg b/setup.cfg
@@ -18,7 +18,7 @@ install_requires =
     torchmetrics >= 1.1.2
     performer-pytorch == 1.1.4
     local-attention == 1.7.1
-    h5py == 3.11.0
+    h5py >= 3.11.0
     numpy >= 1.26.4
     scipy >= 1.13.0
     pandas >= 2.2.2
@@ -29,7 +29,7 @@ install_requires =
     polars >= 0.20.19
     tqdm >= 4.66.2
     pyfaidx == 0.8.1.1
-    gtfparse @ git+https://github.com/TRISTAN-ORF/gtfparse.git#egg=gtfparse
+    gtfparse @ git+https://github.com/TRISTAN-ORF/gtfparse.git@dev#egg=gtfparse
 
 [options.entry_points]
 console_scripts =

diff --git a/transcript_transformer/__init__.py b/transcript_transformer/__init__.py
@@ -65,7 +65,7 @@
     "TGT": "C",
     "TGA": "_",
     "TGG": "W",
-    "NNN": "_",
+    "NNN": "X",
 }
 
 PROT_IDX_DICT = {
@@ -90,6 +90,8 @@
     "W": 18,
     "Y": 19,
     "V": 20,
+    "_": 21,
+    "X": 22,
 }
 
 DNA_IDX_DICT = {

diff --git a/transcript_transformer/data.py b/transcript_transformer/data.py
@@ -15,7 +15,6 @@
 import h5max
 import pyfaidx
 from gtfparse import read_gtf
-
 from .util_functions import vec2DNA, construct_prot, time, slice_gen, prot2vec
 
 REQ_HEADERS = [

diff --git a/transcript_transformer/util_functions.py b/transcript_transformer/util_functions.py
@@ -20,10 +20,10 @@ def construct_prot(seq):
 
     string = ""
     for cdn in cdn_seq:
-        if "N" in cdn:
-            string += "_"
-        else:
+        if cdn in CDN_PROT_DICT.keys():
             string += CDN_PROT_DICT[cdn]
+        else:
+            string += "X"
 
     return string, has_stop, stop_codon