Version history
1 version on record. Newest first; the live version sits at the top with a live indicator.
- Live5/29/2026, 7:45:15 PM
sha256:8e5c3Content snapshot
{ "cells": [ { "source": "# BindCraft Config: CD38 Catalytic-Cleft Binder Design\n\n**Target:** CD38 (UniProt P28907) | **Structure:** PDB 2PGJ (1.71 Å X-ray, chain A)\n\n**Hotspot residues:** C119, T144, D179, N182, C201, S274 (catalytic cleft)\n\n**Baseline comparator:** Apigenin IC50 ~10 µM vs CD38 (Escande et al. J Nutr Biochem 2013)\n\n**Selectivity anti-targets:** CD157/BST1 (UniProt Q10588) — structurally related NADase\n\n**Success criteria:**\n- AF3-Multimer ipTM ≥ 0.6 (independent validation oracle, NOT the design model)\n- Rosetta interface ΔΔG ≤ −1.5 kcal/mol\n- Predicted Kd < 1 µM\n- Foldseek TM-score < 0.5 vs PDB monomer library (novelty gate)\n- ESM-2 log-likelihood perplexity ≤ 8 (sequence naturalness filter)\n\n**Kill criteria (from research plan):**\n- Abort if 2PGJ chain A residue numbering mismatches canonical UniProt P28907 hotspot numbering\n- Abort if AF3-Multimer best design ipTM < 0.4 after 100 designs", "cell_id": "c-61490dd3", "outputs": [], "cell_hash": "sha256:3582e415b21eac57cc33575e6b983d3254836fa9db31be0c09ab925b1985e4fd", "cell_type": "markdown", "execution_count": null }, { "source": "# BindCraft run configuration — CD38 catalytic-cleft binder\n# Reference: Pacesa et al. bioRxiv 2024 (BindCraft)\n# Target structure: PDB 2PGJ, chain A\n\nbindcraft_config = {\n \"target\": {\n \"pdb_id\": \"2PGJ\",\n \"chain_id\": \"A\",\n \"hotspot_residues\": [119, 144, 179, 182, 201, 274],\n \"hotspot_residue_labels\": [\"C119\", \"T144\", \"D179\", \"N182\", \"C201\", \"S274\"],\n \"neighbor_radius_angstrom\": 5.0,\n \"binding_site\": \"catalytic_cleft_NAD_substrate_pocket\"\n },\n \"binder\": {\n \"length_range\": [65, 80],\n \"secondary_structure_bias\": \"mixed_alpha_beta\",\n \"disulfide_allowed\": False,\n \"n_designs\": 100\n },\n \"design_model\": {\n \"backbone_model\": \"RFdiffusion\",\n \"sequence_model\": \"ProteinMPNN\",\n \"mpnn_sampling_temp\": 0.1,\n \"mpnn_sequences_per_backbone\": 8\n },\n \"validation\": {\n \"forward_fold_oracle\": \"AlphaFold2-Multimer\",\n \"independent_oracle\": True,\n \"iptm_threshold\": 0.6,\n \"plddt_threshold\": 80.0,\n \"rosetta_ddg_threshold_kcal_mol\": -1.5\n },\n \"novelty\": {\n \"foldseek_db\": \"pdb100\",\n \"tm_score_max\": 0.5\n },\n \"sequence_filter\": {\n \"esm2_model\": \"esm2_t33_650M_UR50D\",\n \"max_perplexity\": 8.0\n },\n \"selectivity_screen\": {\n \"anti_target_pdb\": \"1ISG\",\n \"anti_target_uniprot\": \"Q10588\",\n \"anti_target_label\": \"CD157_BST1\",\n \"selectivity_fold_min\": 100\n },\n \"baseline_comparator\": {\n \"compound\": \"apigenin\",\n \"ic50_uM\": 10.0,\n \"reference\": \"Escande et al. J Nutr Biochem 2013\",\n \"affinity_type\": \"IC50\"\n },\n \"output\": {\n \"top_n_designs\": 10,\n \"output_format\": [\"pdb\", \"fasta\", \"scores_csv\"]\n }\n}\n\nprint(\"BindCraft config loaded.\")\nprint(f\"Target: PDB {bindcraft_config['target']['pdb_id']} chain {bindcraft_config['target']['chain_id']}\")\nprint(f\"Hotspots: {bindcraft_config['target']['hotspot_residue_labels']}\")\nprint(f\"Binder length: {bindcraft_config['binder']['length_range'][0]}–{bindcraft_config['binder']['length_range'][1]} aa\")\nprint(f\"n_designs: {bindcraft_config['binder']['n_designs']}\")\nprint(f\"Validation oracle: {bindcraft_config['validation']['forward_fold_oracle']} (independent={bindcraft_config['validation']['independent_oracle']})\")\nprint(f\"ipTM threshold: {bindcraft_config['validation']['iptm_threshold']}\")\nprint(f\"Rosetta ΔΔG threshold: {bindcraft_config['validation']['rosetta_ddg_threshold_kcal_mol']} kcal/mol\")", "cell_id": "c-96708d65", "outputs": [], "cell_hash": "sha256:bdbc69412b8a2f6dc2b397bae486570272b425567754b2dab3698ec439a550b1", "cell_type": "code", "execution_count": null } ], "metadata": {}, "owner_ref": "persona-kris-ganjam", "created_by": "persona-kris-ganjam" }