openJiuwen-ai/sciencediscovery / skills/antibody-design
antibody-design
UnverifiedPrepare, launch, monitor, and summarize the real RFdiffusion to ProteinMPNN to Protenix antibody pipeline on a local or remote ScienceDiscovery Runner with sandboxed Ascend NPUs.
| File | Kind | Size |
|---|---|---|
| AGENT_BOOTSTRAP.md | markdown | 1.9 KB |
| SKILL.md | skill md | 11.5 KB |
| references/real_pipeline_config.example.json | config | 506 B |
| requirements.txt | manifest | 863 B |
| scripts/antibody_pipeline_manager.py | script | 47.8 KB |
| scripts/check_real_pipeline_env.sh | script | 11.3 KB |
| scripts/install_hmmer_for_protenix_msa.sh | script | 4.6 KB |
| scripts/install_pipeline_deps.sh | script | 7.7 KB |
| scripts/pdb_to_protenix_json.py | script | 4.0 KB |
| scripts/protenix_py312_compat.py | script | 2.0 KB |
| scripts/rfdiffusion_mindspore_io.patch | other | 7.4 KB |
| scripts/run_after_rfdiffusion.sh | script | 20.4 KB |
| scripts/run_full_antibody_pipeline.sh | script | 17.1 KB |
| scripts/run_sandbox_pipeline.sh | script | 2.3 KB |
| scripts/sandbox_npu_runtime.sh | script | 1.4 KB |
| scripts/screen_protenix_results.py | script | 18.3 KB |
| scripts/test_antibody_pipeline_manager.py | script | 20.3 KB |
| scripts/test_validate_managed_environment.py | script | 3.1 KB |
| scripts/validate_managed_environment.py | script | 4.2 KB |
scripts/rfdiffusion_mindspore_io.patch
View on GitHubdiff --git a/MindSPONGE/applications/rf_diffusion/rfdiffusion/inference/ab_util.py b/MindSPONGE/applications/rf_diffusion/rfdiffusion/inference/ab_util.py
index 29a676e..6b488f7 100644
--- a/MindSPONGE/applications/rf_diffusion/rfdiffusion/inference/ab_util.py
+++ b/MindSPONGE/applications/rf_diffusion/rfdiffusion/inference/ab_util.py
@@ -21,6 +21,7 @@ import logging
from typing import List, Optional
import mindspore as ms
+import numpy as np
from ..kinematics import xyz_to_t2d
from ..parsers import *
@@ -657,107 +658,111 @@ def ab_write_pdblines(
A dictionary mapping loop names to lists of residue indices
"""
- ctr = 1
+ def _to_numpy(value):
+ if value is None:
+ return None
+ if isinstance(value, ms.Tensor):
+ return np.asarray(value.asnumpy())
+ return np.asarray(value)
+
+ atoms = _to_numpy(atoms)
+ seq = np.asarray(_to_numpy(seq)).reshape(-1)
+ chain_idx = _to_numpy(chain_idx)
+ nres = int(seq.shape[0])
if bfacts is None:
- bfacts = ms.mint.zeros(atoms.shape[0])
+ bfacts = np.zeros(nres, dtype=np.float32)
+ else:
+ bfacts = _to_numpy(bfacts)
if idx_pdb is None:
- # Default to 1-indexed residue numbers
- idx_pdb = 1 + ms.mint.arange(atoms.shape[0])
+ idx_pdb = np.arange(1, nres + 1, dtype=np.int32)
+ else:
+ idx_pdb = _to_numpy(idx_pdb)
if loop_map is None:
loop_map = {}
- Bfacts = np.clip(
- bfacts,
- a_min=0,
- a_max=1,
- )
+ Bfacts = np.clip(bfacts, a_min=0, a_max=1)
pdblines = []
- for i in range(seq.shape[0]):
- chain = chain_idx[i]
+ ctr = 1
+ for i in range(nres):
+ c = chain_idx[i]
+ if isinstance(c, (bytes, np.bytes_)):
+ chain = c.decode()
+ else:
+ chain = str(c)
+ seq_i = int(seq[i])
+ res_name = num2aa[seq_i]
+ res_idx = int(idx_pdb[i])
+ bfac = float(Bfacts[i])
# If the input is a single set of atomic coordinates, assume it is a C-alpha trace
if len(atoms.shape) == 2:
-
pdblines.append(
stamp_pdbline(
prefix="ATOM",
ctr=ctr,
atom_name=" CA ",
- residue_name=num2aa[seq[i]],
+ residue_name=res_name,
chain=chain,
- residue_idx=idx_pdb[i],
- x_coord=atoms[i, 0],
- y_coord=atoms[i, 1],
- z_coord=atoms[i, 2],
+ residue_idx=res_idx,
+ x_coord=float(atoms[i, 0]),
+ y_coord=float(atoms[i, 1]),
+ z_coord=float(atoms[i, 2]),
occupancy=1.0,
- b_factor=Bfacts[i],
+ b_factor=bfac,
)
)
-
ctr += 1
# If the input is a set of atomic coordinates with 3 atoms per residue,
# assume it is a backbone trace
elif atoms.shape[1] == 3:
for j, atm_j in enumerate([" N ", " CA ", " C "]):
-
pdblines.append(
stamp_pdbline(
prefix="ATOM",
ctr=ctr,
atom_name=atm_j,
- residue_name=num2aa[seq[i]],
+ residue_name=res_name,
chain=chain,
- residue_idx=idx_pdb[i],
- x_coord=atoms[i, j, 0],
- y_coord=atoms[i, j, 1],
- z_coord=atoms[i, j, 2],
+ residue_idx=res_idx,
+ x_coord=float(atoms[i, j, 0]),
+ y_coord=float(atoms[i, j, 1]),
+ z_coord=float(atoms[i, j, 2]),
occupancy=1.0,
- b_factor=Bfacts[i],
+ b_factor=bfac,
)
)
-
ctr += 1
# If the input is a set of atomic coordinates with 4 atoms per residue,
# assume it is a backbone trace with an oxygen atom
elif atoms.shape[1] == 4:
for j, atm_j in enumerate([" N ", " CA ", " C ", " O "]):
-
pdblines.append(
stamp_pdbline(
prefix="ATOM",
ctr=ctr,
atom_name=atm_j,
- residue_name=num2aa[seq[i]],
+ residue_name=res_name,
chain=chain,
- residue_idx=idx_pdb[i],
- x_coord=atoms[i, j, 0],
- y_coord=atoms[i, j, 1],
- z_coord=atoms[i, j, 2],
+ residue_idx=res_idx,
+ x_coord=float(atoms[i, j, 0]),
+ y_coord=float(atoms[i, j, 1]),
+ z_coord=float(atoms[i, j, 2]),
occupancy=1.0,
- b_factor=Bfacts[i],
+ b_factor=bfac,
)
)
-
ctr += 1
# Otherwise, assume the input is a full atomic tensor with either 14 or 27 atoms per residue
else:
natoms = atoms.shape[1]
+ assert natoms in (14, 27), "Invalid number of atoms per residue, must be 14 or 27"
+ atms = aa2long[seq_i]
- assert (
- natoms in (14, 27)
- ), "Invalid number of atoms per residue, must be 14 or 27"
-
- atms = aa2long[aa2num[seq[i]]]
-
- # his prot hack
- if (
- aa2num[seq[i]] == 8
- and ms.ops.norm(atoms[i, 9, :] - atoms[i, 5, :]) < 1.7
- ):
+ if seq_i == 8 and np.linalg.norm(atoms[i, 9, :] - atoms[i, 5, :]) < 1.7:
atms = (
" N ",
" CA ",
@@ -795,14 +800,14 @@ def ab_write_pdblines(
prefix="ATOM",
ctr=ctr,
atom_name=atm_j,
- residue_name=seq[i],
+ residue_name=res_name,
chain=chain,
- residue_idx=idx_pdb[i],
- x_coord=atoms[i, j, 0],
- y_coord=atoms[i, j, 1],
- z_coord=atoms[i, j, 2],
+ residue_idx=res_idx,
+ x_coord=float(atoms[i, j, 0]),
+ y_coord=float(atoms[i, j, 1]),
+ z_coord=float(atoms[i, j, 2]),
occupancy=1.0,
- b_factor=Bfacts[i],
+ b_factor=bfac,
)
)
ctr += 1
@@ -834,10 +839,10 @@ def stamp_pdbline(
) -> str:
return (
f"{prefix:<6}"
- f"{ctr:>5} "
+ f"{int(ctr):>5} "
f"{atom_name:>4} "
f"{residue_name:>3} "
- f"{chain}{residue_idx:>4} "
- f"{x_coord:8.3f}{y_coord:8.3f}{z_coord:8.3f}"
- f"{occupancy:6.2f}{b_factor:6.2f}\n"
+ f"{chain}{int(residue_idx):>4} "
+ f"{float(x_coord):8.3f}{float(y_coord):8.3f}{float(z_coord):8.3f}"
+ f"{float(occupancy):6.2f}{float(b_factor):6.2f}\n"
)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |