diff --git a/biomni_env/fixed_env.yml b/biomni_env/fixed_env.yml
index 6b928433d..c324947ee 100644
--- a/biomni_env/fixed_env.yml
+++ b/biomni_env/fixed_env.yml
@@ -739,3 +739,4 @@ dependencies:
       - zarr==2.18.7
       - zict==3.0.0
       - zstandard==0.23.0
+      - mi-googlesearch-python==1.3.0.post1
diff --git a/biomni_env/new_software_v006.sh b/biomni_env/new_software_v006.sh
index 7edf433f2..499a424b9 100644
--- a/biomni_env/new_software_v006.sh
+++ b/biomni_env/new_software_v006.sh
@@ -5,3 +5,4 @@ pip install lazyslide
 pip install "git+https://github.com/YosefLab/popV.git@refs/pull/100/head"
 pip install pybiomart
 pip install fair-esm
+pip install mi-googlesearch-python
diff --git a/tests/test_glycoengineering.py b/tests/test_glycoengineering.py
deleted file mode 100644
index 06dd5f76e..000000000
--- a/tests/test_glycoengineering.py
+++ /dev/null
@@ -1,88 +0,0 @@
-import importlib.util
-import re
-import types
-import unittest
-from pathlib import Path
-
-
-def load_glyco_module() -> types.ModuleType:
-    """Load biomni/tool/glycoengineering.py without importing the biomni package.
-    This avoids package-level imports (e.g., pandas) during test discovery.
-    """
-    repo_root = Path(__file__).resolve().parents[1]
-    mod_path = repo_root / "biomni" / "tool" / "glycoengineering.py"
-    spec = importlib.util.spec_from_file_location("biomni_tool_glycoengineering", str(mod_path))
-    module = importlib.util.module_from_spec(spec)
-    assert spec.loader is not None
-    spec.loader.exec_module(module)
-    return module
-
-
-class TestGlycoengineering(unittest.TestCase):
-    @classmethod
-    def setUpClass(cls):
-        cls.mod = load_glyco_module()
-
-    def test_find_n_glycosylation_motifs_basic_and_overlap(self):
-        f = self.mod.find_n_glycosylation_motifs
-
-        # Single canonical sequon
-        out = f("NVT")
-        self.assertIn("Total sequons found: 1", out)
-        self.assertRegex(out, r"-\s+1:\s+NVT")
-
-        # X cannot be Proline
-        out = f("NPT")
-        self.assertIn("Total sequons found: 0", out)
-
-        # Overlapping sequons: NNS (pos1) and NST (pos2)
-        seq = "NNST"
-        out_no_overlap = f(seq, allow_overlap=False)
-        self.assertIn("Total sequons found: 1", out_no_overlap)
-        self.assertRegex(out_no_overlap, r"-\s+1:\s+NNS")
-        self.assertNotRegex(out_no_overlap, r"-\s+2:\s+NST")
-
-        out_overlap = f(seq, allow_overlap=True)
-        self.assertIn("Total sequons found: 2", out_overlap)
-        self.assertRegex(out_overlap, r"-\s+1:\s+NNS")
-        self.assertRegex(out_overlap, r"-\s+2:\s+NST")
-
-    def test_predict_o_glycosylation_hotspots_density_and_proline_rule(self):
-        g = self.mod.predict_o_glycosylation_hotspots
-
-        # S/T rich segment should yield candidates
-        out = g("AAAASTTAAAA", window=7, min_st_fraction=0.3)
-        self.assertRegex(out, r"Total candidate sites: \d+")
-        # Should list at least one 'pos N (' line
-        self.assertRegex(out, r"-\s+pos\s+\d+\s+\([ST]\):")
-
-        # Disallow S/T immediately followed by Proline
-        out2 = g("A\nSP\nASTPA\n".replace("\n", ""), window=3, min_st_fraction=0.34, disallow_proline_next=True)
-        # Ensure no listing for positions where S/T is followed by P
-        # The log lists lines like: - pos N (S): ...
-        lines = [ln.strip() for ln in out2.splitlines() if ln.strip().startswith("- pos ")]
-        for ln in lines:
-            m = re.search(r"- pos (\d+) \(([ST])\):", ln)
-            if not m:
-                continue
-            pos = int(m.group(1))
-            m.group(2)
-            seq = "ASP ASTPA".replace(" ", "")
-            if pos < len(seq) and seq[pos] == "P":
-                self.fail("Proline-followed S/T site should have been excluded: " + ln)
-
-        # Allowing proline-next should include such positions if density passes
-        out3 = g("ASTPA", window=3, min_st_fraction=0.34, disallow_proline_next=False)
-        self.assertRegex(out3, r"Total candidate sites: \d+")
-
-    def test_list_glycoengineering_resources_contains_expected_links(self):
-        lister = self.mod.list_glycoengineering_resources
-        out = lister()
-        # Check presence of the URLs referenced by the issue
-        self.assertIn("https://gitlab.mpcdf.mpg.de/dioscuri-biophysics/glycoshield-md/", out)
-        self.assertIn("https://github.com/CopenhagenCenterForGlycomics", out)
-        self.assertIn("https://services.healthtech.dtu.dk/services/NetOGlyc-4.0/", out)
-
-
-if __name__ == "__main__":
-    unittest.main()