CPL strong/weak binder structures + tables + PWMs (tcren2)

.gitattributes

@@ -408,3 +408,5 @@ Native2022/5sws.pdb filter=lfs diff=lfs merge=lfs -text
 Native2022/*.pdb filter=lfs diff=lfs merge=lfs -text
 Tcr3d2026/*.cif filter=lfs diff=lfs merge=lfs -text
 Native2026/*.pdb filter=lfs diff=lfs merge=lfs -text
+cpl/notebooks_cleanup/hla_table_full.tsv filter=lfs diff=lfs merge=lfs -text
+cpl/notebooks_cleanup/vdjdb_full.txt filter=lfs diff=lfs merge=lfs -text

cpl/content.md

@@ -0,0 +1,17 @@
+## Statistical potentials, CPL, etc.
+
+This repository contains the structures and notebooks for structures data processing. 
+
+Folder `pdb_cpl` contains TCRmodel2 generated TCR:pMHC structures for the 1e6, 4c6, ila1, mel5, and mel8 TCRs with their best and worst peptides. Additionally these folders contains .tsv tables with TCRmodel2 inputs.
+
+The folder `notebooks_cleanup` contains cleaned up notebooks for structure processing allong with folders with corresponding tables and .svg figures. 
+
+`notebooks_cleanup/[1e6, 4c6, ila1, mel5, mel8]_contacts_clean.ipynb` - notebooks for structures processing. Corresponding folders - `1E6`, `4C6`, `ILA1`, `MEL5`, and `MEL8` contains svg figures and intermediate tables.  Absolute paths to data were avoided if it was possible. 
+
+`notebooks_cleanup/dockq_folder.sh` - bash script for running DockQ on folder with .pdb files.
+
+`notebooks_cleanup/DockQ_output_clean.ipynb` - notebook for DockQ results parsinng and processing.
+
+`notebooks_cleanup/tcr_vdb_cleaned.ipynb` - notebook for tcr_vdb structures processing.
+
+`TCRvdb` - folder with TCRvdb pdb and its statistics

cpl/notebooks_cleanup/.ipynb_checkpoints/Structures_copying-checkpoint.ipynb

@@ -0,0 +1,6 @@
+{
+ "cells": [],
+ "metadata": {},
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

cpl/notebooks_cleanup/.ipynb_checkpoints/content-checkpoint.md

@@ -0,0 +1,15 @@
+## Statistical potentials, CPL, etc.
+
+This repository contains the structures and notebooks for structures data processing. 
+
+Folder `pdb_cpl` contains TCRmodel2 generated TCR:pMHC structures for the 1e6, 4c6, ila1, mel5, and mel8 TCRs with their best and worst peptides. Additionally these folders contains .tsv tables with TCRmodel2 inputs.
+
+The folder `notebooks_cleanup` contains cleaned up notebooks for structure processing allong with folders with corresponding tables and .svg figures. 
+
+`notebooks_cleanup/[1e6, 4c6, ila1, mel5, mel8]_contacts_clean.ipynb` - notebooks for structures processing. Corresponding folders - `1E6`, `4C6`, `ILA1`, `MEL5`, and `MEL8` contains svg figures and intermediate tables.  Absolute paths to data were avoided if it was possible. 
+
+`dockq_folder.sh` - bash script for running DockQ on folder with .pdb files.
+
+`DockQ_output_clean.ipynb` - notebook for DockQ results parsinng and processing.
+
+`tcr_vdb_cleaned.ipynb` - notebook for tcr_vdb structures processing.

cpl/notebooks_cleanup/.ipynb_checkpoints/mel5_contacts_clean-checkpoint.ipynb

@@ -0,0 +1,866 @@
+{
+ "cells": [
+  {
+   "id": "fc1a0487",
+   "metadata": {},
+   "source": [
+    "# MEL5 TCR\u2013pMHC contact analysis\n",
+    "\n",
+    "Pipeline for the **MEL5** TCR (HLA-A*02:01 restricted, 10-mer peptides; index\n",
+    "epitope `ELAGIGILTV`). It analyses TCR\u2013peptide\u2013MHC contacts across the\n",
+    "AlphaFold-modelled best- and worst-binding peptide-swap structures.\n",
+    "\n",
+    "**Outputs.** Everything this notebook saves is written under a single `mel5/` folder:\n",
+    "\n",
+    "- `mel5/figures/` \u2014 every figure, saved as **SVG**\n",
+    "- `mel5/tables/`  \u2014 important intermediate tables (TSV)\n",
+    "\n",
+    "Two helpers, `save_fig(name)` and `save_table(df, name)` (defined in the setup\n",
+    "section), handle all saving. Paths that begin with `/projects/...` or\n",
+    "`/home/dluppov/...` point at the original input data and may need adjusting."
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "0939e0ec",
+   "metadata": {},
+   "source": [
+    "## 1. Setup and configuration"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "bbcee98a",
+   "metadata": {},
+   "source": [
+    "import os\n",
+    "from pathlib import Path\n",
+    "from collections import defaultdict\n",
+    "import itertools\n",
+    "import random\n",
+    "\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import seaborn as sns\n",
+    "\n",
+    "import tqdm\n",
+    "import logomaker\n",
+    "\n",
+    "from Bio import PDB\n",
+    "from Bio.PDB import PDBParser\n",
+    "from Bio.SeqUtils import seq1\n",
+    "\n",
+    "from scipy.stats import mannwhitneyu, ttest_ind\n",
+    "from sklearn.metrics import roc_auc_score\n",
+    "\n",
+    "# Monospace font keeps single-letter residue labels aligned in the figures.\n",
+    "plt.rcParams[\"font.family\"] = \"monospace\"\n",
+    "plt.rcParams[\"font.monospace\"] = [\"Courier New\"] + plt.rcParams[\"font.monospace\"]"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "4119768c",
+   "metadata": {},
+   "source": [
+    "# ---------------------------------------------------------------------------\n",
+    "# All outputs from this notebook are written under a single project folder.\n",
+    "# ---------------------------------------------------------------------------\n",
+    "OUTPUT_DIR = Path(\"mel5\")\n",
+    "FIG_DIR = OUTPUT_DIR / \"figures\"   # all figures, saved as SVG\n",
+    "TABLE_DIR = OUTPUT_DIR / \"tables\"  # important intermediate tables (TSV)\n",
+    "\n",
+    "for _d in (FIG_DIR, TABLE_DIR):\n",
+    "    _d.mkdir(parents=True, exist_ok=True)\n",
+    "\n",
+    "\n",
+    "def save_fig(name, fig=None, dpi=300):\n",
+    "    \"\"\"Save the current (or given) matplotlib figure as SVG into mel5/figures.\"\"\"\n",
+    "    fig = fig if fig is not None else plt.gcf()\n",
+    "    path = FIG_DIR / f\"{name}.svg\"\n",
+    "    fig.savefig(path, format=\"svg\", bbox_inches=\"tight\", transparent=True, dpi=dpi)\n",
+    "    print(f\"Saved figure: {path}\")\n",
+    "    return path\n",
+    "\n",
+    "\n",
+    "def save_table(df, name, index=True):\n",
+    "    \"\"\"Save an important intermediate table as TSV into mel5/tables.\"\"\"\n",
+    "    path = TABLE_DIR / f\"{name}.tsv\"\n",
+    "    df.to_csv(path, sep=\"\\t\", index=index)\n",
+    "    print(f\"Saved table: {path}\")\n",
+    "    return path"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "8e2a7217",
+   "metadata": {},
+   "source": [
+    "## 2. Core functions"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "cc36dd65",
+   "metadata": {},
+   "source": [
+    "def extract_coords_from_pdb_by_dict(pdb_file, seq_dict):\n",
+    "    \"\"\"Extract CA and all-atom coordinates for the sequences listed in ``seq_dict``.\n",
+    "\n",
+    "    ``seq_dict`` maps an amino-acid sequence (as found in the PDB) to the label\n",
+    "    used for that segment, e.g. ``{\"CAVNVAGKSTF\": \"CDR3_alpha\", ...}``.\n",
+    "    \"\"\"\n",
+    "    parser = PDBParser(QUIET=True)\n",
+    "    structure = parser.get_structure(\"protein\", pdb_file)\n",
+    "\n",
+    "    coords_ca = {}\n",
+    "    coords_all = {}\n",
+    "    atom_to_ca_map = {}\n",
+    "\n",
+    "    for model in structure:\n",
+    "        for chain in model.get_chains():\n",
+    "            residues = [res for res in chain if PDB.Polypeptide.is_aa(res)]\n",
+    "            pdb_sequence = \"\".join(seq1(res.get_resname()) for res in residues)\n",
+    "\n",
+    "            for seq, label in seq_dict.items():\n",
+    "                if seq not in pdb_sequence:\n",
+    "                    continue\n",
+    "                start_idx = pdb_sequence.find(seq)\n",
+    "                for res in residues[start_idx:start_idx + len(seq)]:\n",
+    "                    res_name = seq1(res.get_resname())\n",
+    "                    res_num = res.id[1]\n",
+    "\n",
+    "                    if \"CA\" in res:\n",
+    "                        ca_key = (label, res_name, res_num)\n",
+    "                        coords_ca[ca_key] = res[\"CA\"].coord.tolist()\n",
+    "                    else:\n",
+    "                        ca_key = None\n",
+    "\n",
+    "                    for atom in res.get_atoms():\n",
+    "                        atom_key = (label, res_name, res_num, atom.get_name())\n",
+    "                        coords_all[atom_key] = atom.coord.tolist()\n",
+    "                        if ca_key is not None:\n",
+    "                            atom_to_ca_map[atom_key] = ca_key\n",
+    "\n",
+    "    return coords_ca, coords_all, atom_to_ca_map\n",
+    "\n",
+    "\n",
+    "def calculate_3d_distance(coord1, coord2):\n",
+    "    return np.linalg.norm(np.array(coord1) - np.array(coord2))\n",
+    "\n",
+    "\n",
+    "def find_atomic_contacts(coords_all, atom_to_ca_map, max_distance=5.0,\n",
+    "                         pdb_filename=None, save_dir=None):\n",
+    "    \"\"\"Find inter-chain atomic contacts within ``max_distance`` (Angstrom).\n",
+    "\n",
+    "    Returns the list of contacting atom pairs (residues in 1-letter code), the\n",
+    "    number of residue (CA) pairs in contact and the number of atom pairs in\n",
+    "    contact. If ``pdb_filename`` is given, the contact list is also written out.\n",
+    "    \"\"\"\n",
+    "    atom_contacts_seen = set()\n",
+    "    ca_contacts_seen = set()\n",
+    "    atomic_contacts = []\n",
+    "\n",
+    "    keys_all = list(coords_all.keys())\n",
+    "    for i, (atom_key_i, coord_i) in enumerate(coords_all.items()):\n",
+    "        ca_i = atom_to_ca_map.get(atom_key_i)\n",
+    "        if ca_i is None:\n",
+    "            continue\n",
+    "        for j in range(i + 1, len(keys_all)):\n",
+    "            atom_key_j = keys_all[j]\n",
+    "            ca_j = atom_to_ca_map.get(atom_key_j)\n",
+    "            if ca_j is None or ca_i == ca_j:\n",
+    "                continue\n",
+    "            if ca_i[0] == ca_j[0]:          # same chain -> not an inter-chain contact\n",
+    "                continue\n",
+    "            if calculate_3d_distance(coord_i, coords_all[atom_key_j]) <= max_distance:\n",
+    "                ca_contacts_seen.add(frozenset((ca_i, ca_j)))\n",
+    "                atom_pair = frozenset((atom_key_i, atom_key_j))\n",
+    "                if atom_pair not in atom_contacts_seen:\n",
+    "                    atom_contacts_seen.add(atom_pair)\n",
+    "                    atom_info_i = (atom_key_i[0], ca_i[1], atom_key_i[2], atom_key_i[3])\n",
+    "                    atom_info_j = (atom_key_j[0], ca_j[1], atom_key_j[2], atom_key_j[3])\n",
+    "                    atomic_contacts.append((atom_info_i, atom_info_j))\n",
+    "\n",
+    "    if pdb_filename:\n",
+    "        base_name = os.path.splitext(os.path.basename(pdb_filename))[0]\n",
+    "        contacts_filename = base_name + \"_contacts.txt\"\n",
+    "        if save_dir:\n",
+    "            os.makedirs(save_dir, exist_ok=True)\n",
+    "            contacts_filename = os.path.join(save_dir, contacts_filename)\n",
+    "        with open(contacts_filename, \"w\") as f:\n",
+    "            for a1, a2 in atomic_contacts:\n",
+    "                f.write(f\"{a1} - {a2}\\n\")\n",
+    "        print(f\"Saved atomic-contact list: {contacts_filename}\")\n",
+    "\n",
+    "    return atomic_contacts, len(ca_contacts_seen), len(atom_contacts_seen)"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "7f77771a",
+   "metadata": {},
+   "source": [
+    "def get_peptides_contacts(atomic_contacts):\n",
+    "    \"\"\"Collapse atomic contacts into per-peptide-residue contact sets (1-letter codes).\"\"\"\n",
+    "    peptide_contacts = [c for c in atomic_contacts\n",
+    "                        if c[0][0] == \"peptide\" or c[1][0] == \"peptide\"]\n",
+    "    if not peptide_contacts:\n",
+    "        return pd.DataFrame(columns=[\"chain_from\", \"aa_from\", \"aa_num_from\", \"contacts\"])\n",
+    "\n",
+    "    contacts_dict = defaultdict(set)\n",
+    "    for contact in peptide_contacts:\n",
+    "        peptide_atom = contacting_atom = None\n",
+    "        for atom in contact:\n",
+    "            if atom[0] == \"peptide\":\n",
+    "                peptide_atom = atom\n",
+    "            else:\n",
+    "                contacting_atom = atom\n",
+    "        contacts_dict[peptide_atom[:-1]].add(contacting_atom[:-1])\n",
+    "\n",
+    "    res_df_i = pd.Series(contacts_dict).reset_index()\n",
+    "    res_df_i.columns = [\"chain_from\", \"aa_from\", \"aa_num_from\", \"contacts\"]\n",
+    "    res_df_i.sort_values(\"aa_num_from\", inplace=True)\n",
+    "    return res_df_i\n",
+    "\n",
+    "\n",
+    "def get_tcr_mhc_contacts(atomic_contacts):\n",
+    "    \"\"\"Collect CDR<->MHC atomic contacts as a tidy table (1-letter codes).\"\"\"\n",
+    "    cols = [\"chain_from\", \"aa_from\", \"aa_num_from\", \"chain_to\", \"aa_to\", \"aa_num_to\"]\n",
+    "    pairs = [c for c in atomic_contacts\n",
+    "             if (\"CDR\" in c[0][0] or \"CDR\" in c[1][0]) and \"MHC\" in (c[0][0], c[1][0])]\n",
+    "    if not pairs:\n",
+    "        return pd.DataFrame(columns=cols)\n",
+    "\n",
+    "    df = pd.DataFrame(pairs)\n",
+    "    df[\"chain_from\"] = df[0].apply(lambda x: x[0])\n",
+    "    df[\"aa_from\"] = df[0].apply(lambda x: x[1])\n",
+    "    df[\"aa_num_from\"] = df[0].apply(lambda x: x[2])\n",
+    "    df[\"chain_to\"] = df[1].apply(lambda x: x[0])\n",
+    "    df[\"aa_to\"] = df[1].apply(lambda x: x[1])\n",
+    "    df[\"aa_num_to\"] = df[1].apply(lambda x: x[2])\n",
+    "\n",
+    "    df = df[cols].drop_duplicates()\n",
+    "    df.sort_values(by=[\"chain_from\", \"aa_num_from\"], inplace=True)\n",
+    "    return df\n",
+    "\n",
+    "\n",
+    "def add_abscent_contacts(res_df_i):\n",
+    "    \"\"\"Add empty-contact rows for peptide positions that make no contacts.\"\"\"\n",
+    "    peptide = res_df_i.iloc[0].peptide\n",
+    "    pdb_path = res_df_i.iloc[0].pdb_path\n",
+    "    present = set(res_df_i.aa_num_from)\n",
+    "\n",
+    "    absent = []\n",
+    "    for i, letter in enumerate(peptide):\n",
+    "        pos = i + 1\n",
+    "        if pos not in present:\n",
+    "            absent.append(pd.Series({\n",
+    "                \"chain_from\": \"peptide\",\n",
+    "                \"aa_from\": letter,\n",
+    "                \"aa_num_from\": pos,\n",
+    "                \"contacts\": set(),\n",
+    "                \"peptide\": peptide,\n",
+    "                \"pdb_path\": pdb_path,\n",
+    "            }))\n",
+    "\n",
+    "    if absent:\n",
+    "        res_df_i = pd.concat([pd.concat(absent, axis=1).T, res_df_i]).sort_values(\"aa_num_from\")\n",
+    "    return res_df_i"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "30d2a9cd",
+   "metadata": {},
+   "source": [
+    "## 3. Sequence definitions"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "c9e49761",
+   "metadata": {},
+   "source": [
+    "# Index (native) epitope and MHC restriction.\n",
+    "INDEX_PEPTIDES = [\"ELAGIGILTV\"]\n",
+    "PEPTIDE_LEN = 10\n",
+    "\n",
+    "# MEL5 CDR loops used for contact extraction, and the HLA-A*02:01 alpha-chain sequence.\n",
+    "cdr3a = \"CAVNVAGKSTF\"\n",
+    "cdr3b = \"CAWSETGLGTGELFF\"\n",
+    "cdr1a, cdr2a = \"DRGSQS\", \"IYSNGD\"\n",
+    "cdr1b, cdr2b = \"GTSNPN\", \"SVGIG\"\n",
+    "mhc_seq = \"GSHSMRYFFTSVSRPGRGEPRFIAVGYVDDTQFVRFDSDAASQRMEPRAPWIEQEGPEYWDGETRKVKAHSQTHRVDLGTLRGYYNQSEAGSHTVQRMYGCDVGSDWRFLRGYHQYAYDGKDYIALKEDLRSWTAADMAAQTTKHKWEAAHVAEQLRAYLEGTCVEWLRRYLENGKETLQ\"\n",
+    "\n",
+    "\n",
+    "def build_seq_dict(peptide):\n",
+    "    \"\"\"sequence -> label map consumed by extract_coords_from_pdb_by_dict.\"\"\"\n",
+    "    return {\n",
+    "        cdr3a: \"CDR3_alpha\",\n",
+    "        cdr3b: \"CDR3_beta\",\n",
+    "        cdr1a: \"CDR1_alpha\",\n",
+    "        cdr2a: \"CDR2_alpha\",\n",
+    "        cdr1b: \"CDR1_beta\",\n",
+    "        cdr2b: \"CDR2_beta\",\n",
+    "        mhc_seq: \"MHC\",\n",
+    "        peptide: \"peptide\",\n",
+    "    }\n",
+    "\n",
+    "\n",
+    "# Chain label -> sequence, and the residue number at which each segment starts in the PDB.\n",
+    "chain_seqs = {\n",
+    "    \"CDR3_alpha\": cdr3a,\n",
+    "    \"CDR3_beta\": cdr3b,\n",
+    "    \"CDR1_alpha\": cdr1a,\n",
+    "    \"CDR2_alpha\": cdr2a,\n",
+    "    \"CDR1_beta\": cdr1b,\n",
+    "    \"CDR2_beta\": cdr2b,\n",
+    "    \"MHC\": mhc_seq,\n",
+    "}\n",
+    "start_dicts = {\"CDR3_alpha\": 88, \"CDR3_beta\": 91, \"CDR1_alpha\": 26,\n",
+    "               \"CDR2_alpha\": 49, \"CDR1_beta\": 26, \"CDR2_beta\": 47, \"MHC\": 0}"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "60df3387",
+   "metadata": {},
+   "source": [
+    "## 4. Single-structure example (sanity check)"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "b68519b7",
+   "metadata": {},
+   "source": [
+    "# Worked example on a single modelled structure: extract coordinates and find contacts.\n",
+    "example_dir = \"/projects/structures/peptide_swap_mel5\"\n",
+    "example_peptide = sorted(os.listdir(example_dir))[0]\n",
+    "example_pdb = f\"{example_dir}/{example_peptide}/{example_peptide}_pmhc_oc/ranked_0.pdb\"\n",
+    "\n",
+    "_, coords_all, atom_to_ca_map = extract_coords_from_pdb_by_dict(example_pdb, build_seq_dict(example_peptide))\n",
+    "atomic_contacts, n_ca, n_atom = find_atomic_contacts(coords_all, atom_to_ca_map, max_distance=5.0)\n",
+    "\n",
+    "print(f\"{example_peptide}: {n_ca} residue contacts, {n_atom} atomic contacts\")\n",
+    "get_peptides_contacts(atomic_contacts).head()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "8b2a9b6f",
+   "metadata": {},
+   "source": [
+    "## 5. Batch contact extraction \u2014 best peptides"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "c8f25536",
+   "metadata": {},
+   "source": [
+    "def extract_contacts_for_dir(structures_dir):\n",
+    "    \"\"\"Extract peptide contacts and CDR<->MHC contacts for every structure in a directory.\"\"\"\n",
+    "    res_list, tcr_mhc_list = [], []\n",
+    "    for peptide in tqdm.tqdm(os.listdir(structures_dir)):\n",
+    "        pdb_path = f\"{structures_dir}/{peptide}/{peptide}_pmhc_oc/ranked_0.pdb\"\n",
+    "        if not os.path.exists(pdb_path):\n",
+    "            print(\"missing:\", pdb_path)\n",
+    "            continue\n",
+    "\n",
+    "        seq_dict_local = build_seq_dict(peptide)\n",
+    "        _, coords_all, atom_to_ca_map = extract_coords_from_pdb_by_dict(pdb_path, seq_dict_local)\n",
+    "        atomic_contacts, _, _ = find_atomic_contacts(coords_all, atom_to_ca_map, max_distance=5.0)\n",
+    "\n",
+    "        res_df_i = get_peptides_contacts(atomic_contacts)\n",
+    "        if len(res_df_i):\n",
+    "            res_df_i[\"peptide\"] = peptide\n",
+    "            res_df_i[\"pdb_path\"] = pdb_path\n",
+    "            res_df_i = add_abscent_contacts(res_df_i)\n",
+    "            res_df_i = res_df_i[[\"chain_from\", \"aa_from\", \"aa_num_from\", \"peptide\", \"pdb_path\", \"contacts\"]]\n",
+    "            res_list.append(res_df_i)\n",
+    "\n",
+    "        tcr_mhc_contacts = get_tcr_mhc_contacts(atomic_contacts)\n",
+    "        if len(tcr_mhc_contacts):\n",
+    "            tcr_mhc_contacts[\"peptide\"] = peptide\n",
+    "            tcr_mhc_contacts[\"pdb_path\"] = pdb_path\n",
+    "            tcr_mhc_list.append(tcr_mhc_contacts)\n",
+    "    return res_list, tcr_mhc_list\n",
+    "\n",
+    "\n",
+    "def explode_contacts(res_list):\n",
+    "    \"\"\"Concatenate per-structure contact frames and explode them to one row per contact.\"\"\"\n",
+    "    res_df = pd.concat(res_list).reset_index(drop=True)\n",
+    "    res_df.contacts = res_df.contacts.apply(list)\n",
+    "    exploded = res_df.explode(\"contacts\").dropna(subset=[\"contacts\"])\n",
+    "    exploded[\"chain_to\"] = exploded.contacts.apply(lambda x: x[0])\n",
+    "    exploded[\"aa_to\"] = exploded.contacts.apply(lambda x: x[1])\n",
+    "    exploded[\"aa_num_to\"] = exploded.contacts.apply(lambda x: x[2])\n",
+    "    exploded = exploded.reset_index(drop=True)\n",
+    "    exploded.index.name = \"index\"\n",
+    "    return exploded"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "265eb33e",
+   "metadata": {},
+   "source": [
+    "# Extract contacts for the best-binding modelled structures.\n",
+    "res_list, tcr_mhc_list = extract_contacts_for_dir(\"/projects/structures/peptide_swap_mel5\")\n",
+    "\n",
+    "best_contacts = explode_contacts(res_list)\n",
+    "best_contacts[\"is_native\"] = best_contacts.peptide.isin(INDEX_PEPTIDES)\n",
+    "save_table(best_contacts, \"all_peptides_contacts_best\")\n",
+    "best_contacts.head()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "ab62c11a",
+   "metadata": {},
+   "source": [
+    "# CDR<->MHC contacts table for the best-binding peptides.\n",
+    "tcr_mhc_best = pd.concat(tcr_mhc_list).reset_index(drop=True)\n",
+    "tcr_mhc_best[\"is_native\"] = tcr_mhc_best.peptide.isin(INDEX_PEPTIDES)\n",
+    "tcr_mhc_best = tcr_mhc_best[\n",
+    "    [\"chain_from\", \"aa_from\", \"aa_num_from\", \"peptide\", \"is_native\",\n",
+    "     \"pdb_path\", \"chain_to\", \"aa_to\", \"aa_num_to\"]\n",
+    "]\n",
+    "tcr_mhc_best.index.name = \"index\"\n",
+    "save_table(tcr_mhc_best, \"all_tcr_mhc_contacts_best\")\n",
+    "tcr_mhc_best.head()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "42f31824",
+   "metadata": {},
+   "source": [
+    "## 6. Contact heatmaps \u2014 best peptides"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "5320fe7a",
+   "metadata": {},
+   "source": [
+    "def plot_contacts_heatmap(cur_chain, contacts, starts, n_pep,\n",
+    "                          peptide_len=PEPTIDE_LEN, name_prefix=\"\", save=True):\n",
+    "    \"\"\"Contact-frequency heatmap (peptide position x chain position) for one chain.\"\"\"\n",
+    "    cur_seq = chain_seqs[cur_chain]\n",
+    "    chain_start = starts[cur_chain]\n",
+    "\n",
+    "    contacts_cur = contacts[contacts.chain_to == cur_chain].copy()\n",
+    "    contacts_cur[\"aa_num_to_adj\"] = contacts_cur.aa_num_to - chain_start\n",
+    "    contacts_cur[\"contacting_pair\"] = (\n",
+    "        contacts_cur.aa_num_from.astype(str) + \"-\" + contacts_cur.aa_num_to_adj.astype(str)\n",
+    "    )\n",
+    "\n",
+    "    contacts_df = contacts_cur.contacting_pair.value_counts().reset_index()\n",
+    "    contacts_df.columns = [\"contacting_pair\", \"count\"]\n",
+    "    contacts_df[\"num_from\"] = contacts_df.contacting_pair.apply(lambda s: int(s.split(\"-\")[0]))\n",
+    "    contacts_df[\"num_to\"] = contacts_df.contacting_pair.apply(lambda s: int(s.split(\"-\")[1]))\n",
+    "    contacts_df.contacting_pair = list(zip(contacts_df.num_from, contacts_df.num_to))\n",
+    "\n",
+    "    # Fill in every position pair that was never observed with a count of 0.\n",
+    "    all_pairs = set(itertools.product(range(1, peptide_len + 1), range(1, len(cur_seq) + 1)))\n",
+    "    missing = all_pairs.difference(contacts_df.contacting_pair)\n",
+    "    missing_df = pd.DataFrame(\n",
+    "        [{\"contacting_pair\": p, \"count\": 0, \"num_from\": p[0], \"num_to\": p[1]} for p in missing]\n",
+    "    )\n",
+    "    full_df = pd.concat([contacts_df, missing_df], ignore_index=True)\n",
+    "\n",
+    "    pivoted_df = full_df.pivot(index=\"num_to\", columns=\"num_from\", values=\"count\") / n_pep\n",
+    "\n",
+    "    plt.figure(figsize=(len(cur_seq) / 2, 9 / 2))\n",
+    "    sns.heatmap(pivoted_df.T, cbar_kws={\"label\": \"frequency\", \"shrink\": 0.5}, square=True,\n",
+    "                cbar=False, vmin=0, vmax=1, cmap=\"gray_r\", linewidths=0.5, linecolor=\"lightgray\")\n",
+    "    plt.ylabel(\"Peptide position\")\n",
+    "    plt.xlabel(\"Chain position\")\n",
+    "    plt.title(cur_chain.replace(\"_\", \" \"))\n",
+    "    if save:\n",
+    "        save_fig(f\"heatmap_{name_prefix}{cur_chain}\")\n",
+    "    plt.show()\n",
+    "    return pivoted_df"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "3f70d4ab",
+   "metadata": {},
+   "source": [
+    "n_peptide_best = best_contacts.peptide.nunique()\n",
+    "\n",
+    "# Plot each chain and collect the per-chain contact-frequency matrices (CDRs only).\n",
+    "contacts_freqs = []\n",
+    "for chain in chain_seqs:\n",
+    "    pivoted = plot_contacts_heatmap(chain, best_contacts, start_dicts, n_peptide_best)\n",
+    "    if chain != \"MHC\":\n",
+    "        df = pivoted.T\n",
+    "        df.columns = [f\"{col}_{chain}\" for col in df.columns]\n",
+    "        contacts_freqs.append(df.copy())\n",
+    "\n",
+    "contacts_freqs_df = pd.concat(contacts_freqs, axis=1)\n",
+    "save_table(contacts_freqs_df, \"mel5_contacts_freqs_for_heatmap\")\n",
+    "contacts_freqs_df"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "3831b6bd",
+   "metadata": {},
+   "source": [
+    "## 7. MHC total contacts"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "9ed51a5b",
+   "metadata": {},
+   "source": [
+    "# Total (summed) MHC contacts per peptide position, max-normalised.\n",
+    "cur_chain = \"MHC\"\n",
+    "cur_seq = chain_seqs[cur_chain]\n",
+    "chain_start = start_dicts[cur_chain]\n",
+    "\n",
+    "contacts_cur = best_contacts[best_contacts.chain_to == cur_chain].copy()\n",
+    "contacts_cur[\"aa_num_to_adj\"] = contacts_cur.aa_num_to - chain_start\n",
+    "contacts_cur[\"contacting_pair\"] = (\n",
+    "    contacts_cur.aa_num_from.astype(str) + \"-\" + contacts_cur.aa_num_to_adj.astype(str)\n",
+    ")\n",
+    "\n",
+    "contacts_df = contacts_cur.contacting_pair.value_counts().reset_index()\n",
+    "contacts_df.columns = [\"contacting_pair\", \"count\"]\n",
+    "contacts_df[\"num_from\"] = contacts_df.contacting_pair.apply(lambda s: int(s.split(\"-\")[0]))\n",
+    "contacts_df[\"num_to\"] = contacts_df.contacting_pair.apply(lambda s: int(s.split(\"-\")[1]))\n",
+    "contacts_df.contacting_pair = list(zip(contacts_df.num_from, contacts_df.num_to))\n",
+    "\n",
+    "all_pairs = set(itertools.product(range(1, PEPTIDE_LEN + 1), range(1, len(cur_seq) + 1)))\n",
+    "missing = all_pairs.difference(contacts_df.contacting_pair)\n",
+    "missing_df = pd.DataFrame(\n",
+    "    [{\"contacting_pair\": p, \"count\": 0, \"num_from\": p[0], \"num_to\": p[1]} for p in missing]\n",
+    ")\n",
+    "full_df = pd.concat([contacts_df, missing_df], ignore_index=True)\n",
+    "pivoted_df = full_df.pivot(index=\"num_to\", columns=\"num_from\", values=\"count\")\n",
+    "\n",
+    "total_per_position = pd.DataFrame(pivoted_df.sum()) / pivoted_df.sum().max()\n",
+    "\n",
+    "plt.figure(figsize=(1 / 2, PEPTIDE_LEN / 2))\n",
+    "sns.heatmap(total_per_position,\n",
+    "            cbar_kws={\"label\": \"Total contacts, max normalized\", \"pad\": 0.3, \"shrink\": 3},\n",
+    "            square=True, cmap=\"gray_r\", linewidths=0.5, linecolor=\"lightgray\", vmin=0, vmax=1)\n",
+    "plt.ylabel(\"peptide position\")\n",
+    "plt.xlabel(\"Total contacts\")\n",
+    "plt.xticks([], [])\n",
+    "plt.title(cur_chain)\n",
+    "save_fig(\"heatmap_MHC_total_contacts\")\n",
+    "plt.show()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "e542b435",
+   "metadata": {},
+   "source": [
+    "## 8. CPL entropy\n",
+    "\n",
+    "Loads the CPL (combinatorial peptide library) readout for MEL5. The original\n",
+    "notebook never defined ``data_cpl`` here, so this load was added; verify the\n",
+    "column name matches your spreadsheet."
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "ba421267",
+   "metadata": {},
+   "source": [
+    "# CPL MIP-1beta readout for MEL5, pivoted to position x residue.\n",
+    "TCR_COLUMN = \"MEL5 MIP1\ud835\udf37\"\n",
+    "data_raw = pd.read_excel(\"./CPL MEL8 1E6 MEL5 and 4C6.xlsx\")\n",
+    "data_raw[\"residue\"] = data_raw.SubLibrary.apply(lambda x: x[-1])\n",
+    "data_raw[\"position\"] = data_raw.SubLibrary.apply(lambda x: int(x[0]) if len(x) == 2 else int(x[:2]))\n",
+    "data_cpl = data_raw.pivot(index=\"position\", columns=\"residue\", values=TCR_COLUMN)\n",
+    "data_cpl"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "174cc7d1",
+   "metadata": {},
+   "source": [
+    "# CPL data (residue x position); negative readings clipped to zero.\n",
+    "cpl_data = data_cpl.T\n",
+    "cpl_data = cpl_data.applymap(lambda x: 0 if x < 0 else x)\n",
+    "\n",
+    "p_cpl = cpl_data / cpl_data.sum()\n",
+    "entropy = p_cpl.applymap(lambda x: -x * np.log(x) if x > 0 else 0).sum()\n",
+    "save_table(entropy.to_frame(\"entropy\"), \"cpl_entropy\")\n",
+    "\n",
+    "plt.figure(figsize=(1 / 2, 9 / 2))\n",
+    "sns.heatmap(pd.DataFrame(entropy), cbar_kws={\"label\": \"Entropy\", \"shrink\": 0.5},\n",
+    "            square=True, vmin=0, cmap=\"gray\", linewidths=0.5, linecolor=\"lightgray\")\n",
+    "plt.xlabel(\"Entropy\")\n",
+    "plt.xticks([], [])\n",
+    "plt.ylabel(\"Peptide position\")\n",
+    "save_fig(\"cpl_entropy_heatmap\")\n",
+    "plt.show()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "222d0d18",
+   "metadata": {},
+   "source": [
+    "## 9. Index peptide MIP-1\u03b2 profile"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "75727e48",
+   "metadata": {},
+   "source": [
+    "index_pept = INDEX_PEPTIDES[0]   # ELAGIGILTV\n",
+    "\n",
+    "# Per-position MIP-1beta for the index peptide's residue, max-normalised.\n",
+    "res_dict = {}\n",
+    "for pos_num, (_, pos) in enumerate(data_cpl.iterrows(), start=1):\n",
+    "    res_dict[pos_num] = pos[index_pept[pos_num - 1]] / pos.max()\n",
+    "res_ser = pd.Series(res_dict).apply(lambda x: 0 if x < 0 else x)\n",
+    "save_table(res_ser.to_frame(\"mip1b_norm\"), \"index_peptide_mip1b\")\n",
+    "\n",
+    "plt.figure(figsize=(1 / 2, 10 / 2))\n",
+    "sns.heatmap(pd.DataFrame(res_ser), cbar_kws={\"label\": \"MIP-1\\u03b2 values, max normalized\", \"shrink\": 0.5},\n",
+    "            square=True, vmin=0, cmap=\"gray_r\", linewidths=0.5, linecolor=\"lightgray\")\n",
+    "plt.xlabel(\"MIP-1\\u03b2 values\")\n",
+    "plt.xticks([], [])\n",
+    "ticks = [0.5, 1.5, 2.5, 3.5, 4.5, 5.5, 6.5, 7.5, 8.5, 9.5][::-1]\n",
+    "labels = [f\"{i + 1}\\n{index_pept[i]}\" for i in range(len(index_pept))][::-1]\n",
+    "plt.yticks(ticks, labels)\n",
+    "plt.ylabel(\"Peptide position\")\n",
+    "save_fig(\"index_peptide_mip1b_heatmap\")\n",
+    "plt.show()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "563fed3f",
+   "metadata": {},
+   "source": [
+    "## 10. Batch contact extraction \u2014 worst peptides"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "248a60c5",
+   "metadata": {},
+   "source": [
+    "# Extract contacts for the worst-binding modelled structures.\n",
+    "res_list_worst, tcr_mhc_list_worst = extract_contacts_for_dir(\"/projects/structures/peptide_swap_mel5_worst\")\n",
+    "\n",
+    "worst_contacts = explode_contacts(res_list_worst)\n",
+    "save_table(worst_contacts, \"all_peptides_contacts_worst\")\n",
+    "print(worst_contacts.pdb_path.nunique(), \"worst structures\")\n",
+    "worst_contacts.head()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "22a3b606",
+   "metadata": {},
+   "source": [
+    "## 11. TCRen: best vs worst contacting"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "8ebb5366",
+   "metadata": {},
+   "source": [
+    "tcren_potentials = pd.read_csv(\"/home/dluppov/tcren-ms/tcren_potentials_mine.csv\")\n",
+    "\n",
+    "\n",
+    "def tcren_potential(aa_from, aa_to):\n",
+    "    \"\"\"TCRen statistical potential for a (peptide aa, TCR aa) contact.\"\"\"\n",
+    "    hit = tcren_potentials[\n",
+    "        (tcren_potentials[\"residue.aa.from\"] == aa_to)\n",
+    "        & (tcren_potentials[\"residue.aa.to\"] == aa_from)\n",
+    "    ].TCRen\n",
+    "    return hit.iloc[0] if len(hit) else np.nan\n",
+    "\n",
+    "\n",
+    "# TCRen is computed only for peptide<->TCR contacts (exclude MHC).\n",
+    "contacts_for_tcren = best_contacts[best_contacts.chain_to != \"MHC\"].reset_index(drop=True)\n",
+    "contacts_for_tcren_worst = worst_contacts[worst_contacts.chain_to != \"MHC\"].reset_index(drop=True)\n",
+    "\n",
+    "contacts_for_tcren[\"tcren_potential\"] = [\n",
+    "    tcren_potential(r.aa_from, r.aa_to) for r in contacts_for_tcren.itertuples()\n",
+    "]\n",
+    "contacts_for_tcren_worst[\"tcren_potential\"] = [\n",
+    "    tcren_potential(r.aa_from, r.aa_to) for r in contacts_for_tcren_worst.itertuples()\n",
+    "]\n",
+    "save_table(contacts_for_tcren, \"contacts_for_tcren_best\")\n",
+    "save_table(contacts_for_tcren_worst, \"contacts_for_tcren_worst\")"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "0deb29e3",
+   "metadata": {},
+   "source": [
+    "tcren_res = contacts_for_tcren.groupby(\"peptide\").tcren_potential.sum()\n",
+    "tcren_res_worst = contacts_for_tcren_worst.groupby(\"peptide\").tcren_potential.sum()\n",
+    "\n",
+    "best = pd.DataFrame(tcren_res); best[\"is_best\"] = 1\n",
+    "worst = pd.DataFrame(tcren_res_worst); worst[\"is_best\"] = 0\n",
+    "potentials = pd.concat([best, worst])\n",
+    "save_table(potentials, \"mel5_potentials\")\n",
+    "print(len(tcren_res), \"best peptides;\", len(tcren_res_worst), \"worst peptides\")"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "68fdb697",
+   "metadata": {},
+   "source": [
+    "# Per-structure potentials: best vs worst.\n",
+    "original = list(tcren_res)\n",
+    "generated = list(tcren_res_worst)\n",
+    "data = pd.DataFrame({\n",
+    "    \"value\": original + generated,\n",
+    "    \"category\": [f\"Best contacting\\nn={len(original)}\"] * len(original)\n",
+    "                + [f\"Worst contacting\\nn={len(generated)}\"] * len(generated),\n",
+    "})\n",
+    "plt.figure(figsize=(8, 6))\n",
+    "sns.boxplot(data=data, x=\"category\", y=\"value\", palette=\"pastel\")\n",
+    "plt.xlabel(\"\")\n",
+    "plt.ylabel(\"TCRen potential\")\n",
+    "p_val = mannwhitneyu(original, generated).pvalue\n",
+    "plt.title(f\"Per structure potentials,\\np={p_val:.2e}\")\n",
+    "save_fig(\"tcren_box_best_vs_worst_per_structure\")\n",
+    "plt.show()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "e0ccf9e1",
+   "metadata": {},
+   "source": [
+    "print(\"t-test:\", ttest_ind(list(tcren_res), list(tcren_res_worst)))\n",
+    "\n",
+    "data2auc = pd.concat([\n",
+    "    pd.DataFrame(list(tcren_res), columns=[\"potential\"]).assign(is_best=1),\n",
+    "    pd.DataFrame(list(tcren_res_worst), columns=[\"potential\"]).assign(is_best=0),\n",
+    "])\n",
+    "auc = roc_auc_score(data2auc[\"is_best\"], -data2auc[\"potential\"])\n",
+    "print(f\"ROC AUC = {auc:.4f}\")"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "39b47bf9",
+   "metadata": {},
+   "source": [
+    "## 12. Sequence logos"
+   ],
+   "cell_type": "markdown"
+  },
+  {
+   "id": "cf4494a7",
+   "metadata": {},
+   "source": [
+    "clones = list(tcren_res.index)\n",
+    "mat_df = logomaker.alignment_to_matrix(clones)\n",
+    "logomaker.Logo(mat_df, color_scheme=\"skylign_protein\")\n",
+    "plt.title(\"Best peptides\", fontsize=20)\n",
+    "plt.ylabel(\"Count\")\n",
+    "plt.xticks(range(0, 10), range(1, 11))\n",
+    "save_fig(\"logo_best_peptides\")\n",
+    "plt.show()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "id": "8bf2a7a5",
+   "metadata": {},
+   "source": [
+    "clones = list(tcren_res_worst.index)\n",
+    "mat_df = logomaker.alignment_to_matrix(clones)\n",
+    "logomaker.Logo(mat_df, color_scheme=\"skylign_protein\")\n",
+    "plt.title(\"Worst peptides\", fontsize=20)\n",
+    "plt.ylabel(\"Count\")\n",
+    "plt.xticks(range(0, 10), range(1, 11))\n",
+    "save_fig(\"logo_worst_peptides\")\n",
+    "plt.show()"
+   ],
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python [conda env:.conda-new_analysis_env]",
+   "language": "python",
+   "name": "conda-env-.conda-new_analysis_env-py"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.11"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

cpl/notebooks_cleanup/1E6/tables/1e6_contacts_freqs_for_heatmap.tsv

@@ -0,0 +1,11 @@
+num_from	1_CDR3_alpha	2_CDR3_alpha	3_CDR3_alpha	4_CDR3_alpha	5_CDR3_alpha	6_CDR3_alpha	7_CDR3_alpha	8_CDR3_alpha	9_CDR3_alpha	10_CDR3_alpha	11_CDR3_alpha	12_CDR3_alpha	13_CDR3_alpha	1_CDR3_beta	2_CDR3_beta	3_CDR3_beta	4_CDR3_beta	5_CDR3_beta	6_CDR3_beta	7_CDR3_beta	8_CDR3_beta	9_CDR3_beta	10_CDR3_beta	11_CDR3_beta	12_CDR3_beta	13_CDR3_beta	14_CDR3_beta	15_CDR3_beta	16_CDR3_beta	1_CDR1_alpha	2_CDR1_alpha	3_CDR1_alpha	4_CDR1_alpha	5_CDR1_alpha	6_CDR1_alpha	1_CDR1_beta	2_CDR1_beta	3_CDR1_beta	4_CDR1_beta	5_CDR1_beta	1_CDR2_beta	2_CDR2_beta	3_CDR2_beta	4_CDR2_beta	5_CDR2_beta	6_CDR2_beta
+1	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+2	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+3	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+4	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+5	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+6	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+7	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+8	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+9	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
+10	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

cpl/notebooks_cleanup/1E6/tables/1e6_potentials.tsv

@@ -0,0 +1,146 @@
+peptide	tcren_potential	is_best
+PIDEIHGCVH	2.4217027493580394	0
+PIDEIQGCVE	3.683395658418914	0
+PIDETHNCVE	3.3507766427432615	0
+PIDETHNDVE	1.18601314892187	0
+PIDETHNDVH	3.5823692485439986	0
+PIDETQGCVE	5.0524486153285295	0
+PIDETQGCVH	4.96961223452746	0
+PIDETQGDVE	5.620770296072169	0
+PIDVIHGDVE	2.939117432742732	0
+PIDVIHNCVE	1.4227879580260236	0
+PIDVIHNDVE	3.439065812757903	0
+PIDVIQGCVE	3.238193676661705	0
+PIDVTQGCVE	5.023703854930251	0
+PIDVTQNDVE	3.514256665078716	0
+PIKEIHGCVE	-0.3508725596133602	0
+PIKETQGDVE	1.1922250171950934	0
+PIKETQGDVH	1.3571629495217683	0
+PIKETQNCVE	-1.4651834448782406	0
+PIKETQNCVH	-1.4359345386821278	0
+PIKVIHGCVE	2.6067560936466414	0
+PIKVIQNCVE	3.561681856269784	0
+PIKVIQNDVE	2.6569541707807693	0
+PIKVTHNCVH	0.7208021846795598	0
+PIKVTQGCVH	1.7631769416595384	0
+PIKVTQGDVE	1.4620821014950194	0
+PIKVTQNCVE	1.8176919412263581	0
+PLDEIHGCVE	2.9963659493276276	0
+PLDEIHNCVE	1.67293274232086	0
+PLDEIHNCVH	1.5079948099941851	0
+PLDEIHNDVH	2.2949014018018024	0
+PLDETHGCVH	2.933438226452563	0
+PLDETHNCVH	2.8308562416217575	0
+PLDVIHGDVH	3.922230325715385	0
+PLDVIQGDVE	5.871870612525463	0
+PLDVIQGDVH	4.1444033558276185	0
+PLDVIQNCVH	2.895285362565613	0
+PLDVIQNDVE	3.9744573101934093	0
+PLDVTHGCVH	3.8497280683872135	0
+PLDVTHGDVH	3.157912107034939	0
+PLDVTHNCVH	3.143884347802661	0
+PLDVTQNDVE	2.586603634051376	0
+PLKEIQGCVH	0.7704583087636465	0
+PLKEIQNDVH	0.2664217375618066	0
+PLKETHGCVH	1.61618175363069	0
+PLKETQGCVH	1.0338744874234191	0
+PLKVIHNDVH	0.29700758841577574	0
+PLKVIQGCVH	1.9338952328382057	0
+PLKVIQGDVE	3.7285810929596366	0
+PLKVIQNCVE	-0.05485791824507996	0
+PLKVTHNCVE	2.0642213917936414	0
+PLKVTQNCVH	-0.039491066614363024	0
+PMDEIQGCVE	3.6535993341296207	0
+PMDETQGCVH	4.918611999261196	0
+PMDETQGDVH	5.0454800392522285	0
+PMDVIHNCVH	1.619187739853329	0
+PMDVIHNDVH	1.5042699528609709	0
+PMDVIQNDVE	2.7972093265057545	0
+PMDVIQNDVH	2.684018117351437	0
+PMDVTHGDVH	1.9362421630228783	0
+PMDVTHNCVH	3.0187734668097237	0
+PMDVTQGDVH	5.02449893351979	0
+PMDVTQNDVE	2.652637693132908	0
+PMKEIHNCVE	0.4045872444256977	0
+PMKEIQGCVE	0.024197181010331414	0
+PMKEIQNCVH	-0.13008497628818022	0
+PMKETHNCVH	-0.3438453681322597	0
+PMKETHNDVH	-0.23144890060305368	0
+PMKVIHGDVE	2.692785645207289	0
+PMKVIHGDVH	0.09780611265182039	0
+PMKVIHNCVH	1.3996376091374636	0
+PMKVIHNDVE	0.3295083040860316	0
+PMKVIQNDVE	-0.11503223693171359	0
+PMKVTHNDVH	0.2562840416768547	0
+PMKVTQGCVE	1.6872362196490653	0
+PMKVTQGDVH	1.9917779178888744	0
+QIDEIHGCVE	3.04197023990714	0
+QIDEIHNCVE	1.16664918636756	0
+QIDEIHNCVH	1.0390830151753319	0
+QIDEIHNDVE	1.4065535096391024	0
+QIDEIQGCVH	3.4750640983834207	0
+QIDETHNCVE	2.9662407121745122	0
+QIDETHNCVH	2.6686725923814123	0
+QIDETQGDVE	5.18841840138324	0
+QIDETQNDVE	4.453984289337388	0
+QIDVIQNCVE	1.8476730751007835	0
+QIDVTHGCVE	5.398030665205182	0
+QIDVTHGDVE	3.959530826457527	0
+QIDVTHNDVE	2.412677869328042	0
+QIDVTQNDVH	2.4490605734096467	0
+QIKEIHGCVH	0.11059210912627289	0
+QIKEIHNCVE	-0.7906303608217574	0
+QIKEIHNCVH	1.3985616871936721	0
+QIKETHGDVE	1.4381821380664448	0
+QIKETHNDVH	1.2879778025539361	0
+QIKETQGDVE	1.3499071898038228	0
+QIKETQNCVH	-1.4502225150582055	0
+QIKVIHGDVE	1.144700848921102	0
+QIKVIHNCVE	3.8615012237825814	0
+QIKVIHNCVH	-0.21054219770174115	0
+QIKVIQGCVH	1.5245147830012729	0
+QIKVIQNDVH	-0.3209132665643869	0
+QIKVTHGCVE	2.015492395867177	0
+QIKVTQGDVH	2.5954443784638683	0
+QIKVTQNDVH	0.3921488342815513	0
+QLDEIHGDVH	1.1683736522483938	0
+QLDEIHNCVH	1.9737055809439448	0
+QLDETHGDVE	5.271413577260525	0
+QLDETQGDVE	6.181203311482637	0
+QLDETQNCVE	4.39971745522595	0
+QLDETQNDVH	3.5728143253777533	0
+QLDVIHGCVH	3.9035846006015253	0
+QLDVIQGDVE	4.966545452647522	0
+QLDVIQNCVH	2.9153544008919114	0
+QLDVIQNDVE	4.258888888530427	0
+QLDVTHGDVH	4.919422401999158	0
+QLDVTHNCVE	3.610262440192716	0
+QLKEIHGDVE	0.03584363160047073	0
+QLKETHGCVH	2.5940639149005715	0
+QLKETHNDVH	1.5293939994903343	0
+QLKVIHNCVH	1.8984173914779345	0
+QLKVIQGDVE	2.688451347352709	0
+QLKVIQNDVH	-0.4867698213376626	0
+QLKVTHGCVH	2.8634872227004666	0
+QLKVTHNDVE	1.0383423050549125	0
+QMDEIHGCVE	2.728087268787391	0
+QMDEIHGDVH	1.1683736522483938	0
+QMDEIQGCVH	3.3461193626020918	0
+QMDEIQGDVH	4.383822337271904	0
+QMDETHGCVE	5.284281675435734	0
+QMDETHNCVE	3.0770684122376575	0
+QMDETQNDVE	3.728308929360309	0
+QMDVIHNDVE	1.977000668762165	0
+QMDVIQGCVH	2.7391280844068704	0
+QMDVIQNCVE	3.071764611318037	0
+QMDVTHNDVH	2.5976140658632185	0
+QMKEIHNCVE	1.0034451524330514	0
+QMKETHGCVH	2.155563750584206	0
+QMKETHGDVE	1.6664551009618773	0
+QMKETHGDVH	1.8530436854083951	0
+QMKVIHGCVE	1.6729581081307288	0
+QMKVIHGCVH	1.6597275158477454	0
+QMKVIHNCVH	0.3778754522875994	0
+QMKVIHNDVE	1.0965634187945552	0
+QMKVTHGCVE	3.8994798086809714	0
+QMKVTQGCVH	2.293732754197879	0

cpl/notebooks_cleanup/1E6/tables/contacts_for_tcren_best.tsv

@@ -0,0 +1 @@
+	chain_from	aa_from	aa_num_from	peptide	is_native	pdb_path	chain_to	aa_to	aa_num_to	tcren_potential

cpl/notebooks_cleanup/1E6/tables/cpl_entropy.tsv

@@ -0,0 +1,11 @@
+position	entropy
+1	2.407471641196575
+2	1.6510143491044356
+3	1.7002128156721832
+4	0.3932386634534798
+5	0.7836237553922625
+6	1.110147305477272
+7	1.481531706796957
+8	1.731924383602427
+9	2.4784140195836883
+10	1.4733252403646113

cpl/notebooks_cleanup/1E6/tables/index_peptide_mip1b.tsv

@@ -0,0 +1,11 @@
+	mip1b_norm
+1	0.16247139578300668
+2	0.3915118031075733
+3	0.82196162034519
+4	1.0
+5	1.0
+6	1.0
+7	0.0
+8	0.0
+9	0.34621643610579267
+10	0.0

cpl/notebooks_cleanup/4C6/tables/.ipynb_checkpoints/4c6_potentials-checkpoint.tsv

@@ -0,0 +1,320 @@
+peptide	tcren_potential	is_best
+LWLPLAGIA	2.274011493240409	1
+LWLPLAGIL	2.1469456496459105	1
+LWLPLAGIM	2.1386479665260314	1
+LWLPLAGIV	2.1386479665260314	1
+LWLPLAGIW	2.1386479665260314	1
+LWLPLAGLA	2.228195321300239	1
+LWLPLAGLL	1.5803074278951503	1
+LWLPLAGLM	2.228195321300239	1
+LWLPLAGLV	3.2058851403929634	1
+LWLPLAGLW	0.9743304301056991	1
+LWLPLAPIA	2.127499426155844	1
+LWLPLAPIL	2.983526104404297	1
+LWLPLAPIM	2.135797109275723	1
+LWLPLAPIV	2.1975533719239175	1
+LWLPLAPIW	2.127499426155844	1
+LWLPLAPLA	2.2253444640499302	1
+LWLPLAPLL	1.6873114272490577	1
+LWLPLAPLM	2.2253444640499302	1
+LWLPLAPLV	1.3870332667474368	1
+LWLPLAPLW	1.023320248474045	1
+LWLPLFGIA	1.9350777011864897	1
+LWLPLFGIL	1.7405677357804255	1
+LWLPLFGIM	1.9971095674964834	1
+LWLPLFGIV	1.7405677357804255	1
+LWLPLFGIW	2.3829354707682793	1
+LWLPLFGLA	1.8220930110965532	1
+LWLPLFGLL	1.1742051176914647	1
+LWLPLFGLM	1.7030492469601337	1
+LWLPLFGLV	2.318704642903043	1
+LWLPLFGLW	1.7549770421141986	1
+LWLPLFPIA	1.5619940662369332	1
+LWLPLFPIL	1.5702917493568123	1
+LWLPLFPIM	1.792218280970481	1
+LWLPLFPIV	1.8090578440799	1
+LWLPLFPIW	1.5619940662369332	1
+LWLPLFPLA	1.5814874752430672	1
+LWLPLFPLL	1.0298530741131156	1
+LWLPLFPLM	0.933599581837979	1
+LWLPLFPLV	1.003653527606052	1
+LWLPLFPLW	1.645745658932107	1
+LWLRLAGIA	1.6427692587125287	1
+LWLRLAGIL	1.095423221224725	1
+LWLRLAGIM	1.564693233486376	1
+LWLRLAGIV	1.51570341511803	1
+LWLRLAGIW	1.51570341511803	1
+LWLRLAGLA	1.654240588260583	1
+LWLRLAGLL	1.6052507698922371	1
+LWLRLAGLM	1.6052507698922371	1
+LWLRLAGLV	1.5351968241241638	1
+LWLRLAGLW	1.6052507698922371	1
+LWLRLAPIA	1.5365701683496	1
+LWLRLAPIL	-0.6911145740741516	1
+LWLRLAPIM	1.5045548747478423	1
+LWLRLAPIV	1.5029195514071119	1
+LWLRLAPIW	1.5045548747478423	1
+LWLRLAPLA	1.5941022295220497	1
+LWLRLAPLL	1.5941022295220497	1
+LWLRLAPLM	1.0482635904711317	1
+LWLRLAPLV	1.5941022295220497	1
+LWLRLAPLW	1.9240366477415827	1
+LWLRLFGIA	1.1096011049143444	1
+LWLRLFGIL	1.0239387679632501	1
+LWLRLFGIM	1.039547159146271	1
+LWLRLFGIV	1.3741650160884817	1
+LWLRLFGIW	0.9205033950098517	1
+LWLRLFGLA	1.1290945139204782	1
+LWLRLFGLL	0.7788682657952464	1
+LWLRLFGLM	1.1290945139204782	1
+LWLRLFGLV	1.6957600914950415	1
+LWLRLFGLW	1.1290945139204782	1
+LWLRLFPIA	0.7194436166277532	1
+LWLRLFPIL	0.5187693209356266	1
+LWLRLFPIM	1.1989299872803916	1
+LWLRLFPIV	0.6090303852808743	1
+LWLRLFPIW	0.6090303852808743	1
+LWLRLFPLA	0.6665624464533237	1
+LWLRLFPLL	0.6665624464533237	1
+LWLRLFPLM	0.9905582174368234	1
+LWLRLFPLV	0.6665624464533237	1
+LWLRLFPLW	0.9382510939293672	1
+LWMRLLPLL	1.1621985201131624	1
+LWPPLAGIA	2.068594020757958	1
+LWPPLAGIL	1.9974820808619786	1
+LWPPLAGIM	2.1876377848943775	1
+LWPPLAGIV	2.1876377848943775	1
+LWPPLAGIW	2.146670045984111	1
+LWPPLAGLA	2.0423284236907215	1
+LWPPLAGLL	2.228195321300239	1
+LWPPLAGLM	2.228195321300239	1
+LWPPLAGLV	2.608994001265285	1
+LWPPLAGLW	2.228195321300239	1
+LWPPLAPIA	2.0918913935797376	1
+LWPPLAPIL	2.127499426155844	1
+LWPPLAPIM	2.1355215056139234	1
+LWPPLAPIV	2.1179151555313753	1
+LWPPLAPIW	2.127499426155844	1
+LWPPLAPLA	2.1638323982713965	1
+LWPPLAPLL	3.024194577486327	1
+LWPPLAPLM	1.3853546716238603	1
+LWPPLAPLV	1.023320248474045	1
+LWPPLAPLW	2.5807597992034426	1
+LWPPLFGIA	1.6172994073536977	1
+LWPPLFGIL	1.6188615985200867	1
+LWPPLFGIM	1.609277327895618	1
+LWPPLFGIV	1.609277327895618	1
+LWPPLFGIW	-0.8061571955426897	1
+LWPPLFGLA	1.5979957219474148	1
+LWPPLFGLL	1.708408953294294	1
+LWPPLFGLM	1.6988246826698252	1
+LWPPLFGLV	2.2750745308688574	1
+LWPPLFGLW	1.6680496677154881	1
+LWPPLFPIA	1.4387257378102056	1
+LWPPLFPIL	1.4387257378102056	1
+LWPPLFPIM	1.4146593914921861	1
+LWPPLFPIV	1.4483100084346743	1
+LWPPLFPIW	1.4483100084346743	1
+LWPPLFPLA	1.5378573632088814	1
+LWPPLFPLL	1.5282730925844128	1
+LWPPLFPLM	1.5282730925844128	1
+LWPPLFPLV	1.887969017153194	1
+LWPPLFPLW	1.5282730925844128	1
+LWPRLAGIA	1.407912542734666	1
+LWPRLAGIL	1.2845209311292174	1
+LWPRLAGIM	0.9796102693832813	1
+LWPRLAGIV	1.3998904632765863	1
+LWPRLAGIW	1.1654771669927984	1
+LWPRLAGLA	1.399092042377022	1
+LWPRLAGLL	1.3740682859034248	1
+LWPRLAGLM	1.4193838722827201	1
+LWPRLAGLV	1.4691459881450952	1
+LWPRLAGLW	1.5622244467003805	1
+LWPRLAPIA	1.4842630448421439	1
+LWPRLAPIL	1.1543286266226107	1
+LWPRLAPIM	1.1340367967169123	1
+LWPRLAPIV	0.8820936923950077	1
+LWPRLAPIW	1.1340367967169123	1
+LWPRLAPLA	0.2948187030777843	1
+LWPRLAPLL	1.2438759813968179	1
+LWPRLAPLM	0.9518955040150759	1
+LWPRLAPLV	1.573810399616351	1
+LWPRLAPLW	1.1248322172603986	1
+LWPRLFGIA	0.6770511605734204	1
+LWPRLFGIL	0.4598639155377411	1
+LWPRLFGIM	1.0192553292405726	1
+LWPRLFGIV	0.6690290811153409	1
+LWPRLFGIW	1.0239387679632501	1
+LWPRLFGLA	1.2605244095266348	1
+LWPRLFGLL	0.758576435889548	1
+LWPRLFGLM	1.1088026840147798	1
+LWPRLFGLV	0.758576435889548	1
+LWPRLFGLW	1.1086880898338607	1
+LWPRLFPIA	0.959116970502039	1
+LWPRLFPIL	0.5567232617734181	1
+LWPRLFPIM	0.84870373915516	1
+LWPRLFPIV	0.5887385553751758	1
+LWPRLFPIW	0.84870373915516	1
+LWPRLFPLA	0.9381364997484481	1
+LWPRLFPLL	-0.15861498183167344	1
+LWPRLFPLM	1.2883627478736797	1
+LWPRLFPLV	0.9382510939293672	1
+LWPRLFPLW	0.9382510939293672	1
+FFCCCGESF	2.493788664220773	0
+FFCCCRFRI	6.939565084107656	0
+FFCCCRFSI	1.7594750226944238	0
+FFCSCEERI	1.6484653031152208	0
+FFCSCGERI	2.126183380559794	0
+FFCSCRESL	0.07303804258158111	0
+FFCSSDERI	2.5633455286950095	0
+FFCSSDESF	1.469640725919887	0
+FFCSSEERI	1.5281306650959288	0
+FFIMMDKKI	-1.6174458104318603	0
+FFIMPRKDI	2.638871037270585	0
+FFIQMEKDL	-0.3086320944158666	0
+FFIQMGKKF	0.726873827882694	0
+FFKMMEKKI	-2.529459688965277	0
+FFKMMEMDI	-0.18666712511912908	0
+FFKMMRKDL	2.0723102940713516	0
+FFKMMRMDL	4.037329755322921	0
+FFKMMRMKF	3.89325308110519	0
+FFKMPGKDI	0.11019888439334355	0
+FFKMPGKDL	0.2576596431690612	0
+FFKQMDMKI	0.2789999721586617	0
+FFKQPDMDL	0.5314040109538414	0
+FFSCCDERF	3.2943054179584466	0
+FFSCCDFRL	4.210145755400449	0
+FFSCCDFSL	3.4109660430805224	0
+FFSCCEFSF	2.3713347206289583	0
+FFSCSDFSI	1.6281105129562938	0
+FFSCSEERL	1.9513561620981108	0
+FFSCSRESL	1.929271445432191	0
+FFSCSRFRL	4.705578696463731	0
+FFSSCEESL	-1.3932105761448907	0
+FFSSCEFSL	-1.3849037931725974	0
+FFSSSDERF	1.9838326076153405	0
+FFSSSDERL	1.7791362838978158	0
+FFSSSDFRL	2.195609441559557	0
+FFSSSEFRL	3.0878213488351474	0
+FFSSSEFSI	-0.35260901382966675	0
+FWCCCDFRF	4.738196266930846	0
+FWCCCEERF	3.3686199680728306	0
+FWCCSRFSI	2.206350865070471	0
+FWCSCDERF	0.18199762794764163	0
+FWCSCEFSI	-1.1376388450065322	0
+FWCSCRESF	1.2688561862976617	0
+FWCSSDERL	2.9059754225303926	0
+FWIMMEMDL	0.31461740296517293	0
+FWIMPDKKI	-3.1652246541349918	0
+FWIMPGKDL	0.8117205726703108	0
+FWIQPEKDL	0.24088942769878546	0
+FWIQPEMDL	1.8154536184993304	0
+FWKMMEMKI	0.17524376419506693	0
+FWKMPEKKF	-2.7595956887562916	0
+FWKMPEMKL	-0.8156447313615405	0
+FWKQPEKKF	-1.2589457795373236	0
+FWKQPRMDF	4.061972878768342	0
+FWSCCEFRI	4.393977618714295	0
+FWSCCGESL	3.4128989411484363	0
+FWSCSRFRI	5.4808718514044195	0
+FWSSCRERF	4.453363000021189	0
+FWSSSEFRF	2.962643251638249	0
+FWSSSGESF	0.6176261443765156	0
+FYCCCGERF	6.194589467650578	0
+FYCSCEERI	1.1238837433722904	0
+FYCSSDFRI	3.4028373398966245	0
+FYIMPGMKI	0.4552196587084154	0
+FYKQMEMKI	0.7944972337863621	0
+FYSCCDERF	4.124258530852192	0
+FYSCCDERL	4.124258530852192	0
+FYSCCEFSI	0.5765904661062682	0
+FYSCCGERF	5.323195029462891	0
+FYSCSDERI	2.7273534468214313	0
+FYSSCEERL	2.5103267132461937	0
+FYSSSGESI	-0.6217108935445976	0
+TFCCCEFRF	3.40914477748398	0
+TFCCCEFSF	2.3568947676755663	0
+TFCCCRESF	1.099320413259733	0
+TFCCSEFSF	2.243739546969789	0
+TFCCSRESF	1.8538612364563778	0
+TFCSCDESF	0.38330724708102615	0
+TFCSCDESI	1.6329478777205528	0
+TFCSCGFRI	4.799837489178535	0
+TFCSSEFSL	-0.09170141385863984	0
+TFCSSGESF	-0.14199750733437505	0
+TFIMMDMDF	-0.16351753894495497	0
+TFIMMDMKF	-0.4748613534377927	0
+TFIMMEKKI	-2.6974903424913474	0
+TFIMMGMDI	2.669189762033672	0
+TFIMMRMDL	4.007613186483783	0
+TFIMPEMKF	-0.6683906162746963	0
+TFIQMRMDI	4.592604318307105	0
+TFIQPGKDF	1.7348788322769004	0
+TFKMMEKDL	-1.5488531142232775	0
+TFKMMRKDF	1.9248055212641473	0
+TFKMPDMDL	-0.17345429837538145	0
+TFKQMDKKL	-1.2401658330148744	0
+TFKQMEKDF	0.6116174105115224	0
+TFKQMRMDF	5.165466724661365	0
+TFKQPDMDL	-0.05139511243723836	0
+TFSCCGFRI	6.845257251511432	0
+TFSCCGFSL	4.488275700373962	0
+TFSCSDESI	1.7579587495174291	0
+TFSCSEESL	-0.1870672488874986	0
+TFSCSEFSF	0.8758723989481765	0
+TFSSCGESL	-0.3830319134585849	0
+TFSSSDERL	3.3331680863443207	0
+TFSSSEESL	-0.4709883233914507	0
+TFSSSRERF	2.0333533672055184	0
+TWCCCDFSF	2.215296548851103	0
+TWCCSRESF	0.6849953478212809	0
+TWCSCGESI	-1.280219142533488	0
+TWCSSDFSI	-0.2993566678189006	0
+TWIMPDMDF	-0.2153685922942456	0
+TWIMPGKKL	-1.374380948381484	0
+TWIQMEKKF	-1.5210986347860045	0
+TWIQMEMDI	1.998127510823304	0
+TWIQMRKDL	3.5211947234018752	0
+TWIQPGMKI	2.526960709711327	0
+TWIQPRKDI	1.5988372962854944	0
+TWIQPRKKL	-1.2564044634552984	0
+TWKMMDKDF	-0.6043875619086088	0
+TWSCCRESL	3.1597102407033946	0
+TWSCSGESL	1.2048217822006901	0
+TWSCSRESF	-0.09757002310928853	0
+TWSCSRFSI	2.889250960014654	0
+TWSSCDESF	0.9248682967833317	0
+TWSSCDFRI	2.0584431558585488	0
+TWSSCEFSL	-1.3323777310675537	0
+TWSSCGFSF	1.2061102825319492	0
+TWSSCRERF	3.9059279354367087	0
+TWSSSDERL	3.437615168175161	0
+TWSSSDFRF	2.9016757749783357	0
+TYCCCRFSF	2.38598779092243	0
+TYCCCRFSL	0.0410568227359383	0
+TYCCSDESI	3.432717991992763	0
+TYCSCEFSF	-1.553662958520015	0
+TYCSCGFRI	2.1879690696039704	0
+TYCSCRERI	2.0564698239278725	0
+TYCSCRESF	0.6589104737584217	0
+TYCSSEFRL	3.272344590253676	0
+TYCSSRESL	1.3595479104378054	0
+TYCSSRFRI	2.263803795762274	0
+TYIMMDMDL	0.4376727701086003	0
+TYIMPGMKI	1.4853171003654797	0
+TYIQMDMDL	1.7077658706079655	0
+TYIQMGMKI	1.090618218465069	0
+TYIQPDKKF	-0.9329631593992991	0
+TYKMPRKKF	0.5631213340186528	0
+TYKQMGKDL	2.595764399340909	0
+TYKQMGKKL	0.34737782988448096	0
+TYKQPDKKI	-2.1041632171507882	0
+TYSCCDESI	1.5330509839066786	0
+TYSCCRESL	3.160534998403393	0
+TYSCSGFRL	4.536395468268973	0
+TYSSCEFRL	0.6433060395945056	0
+TYSSCGERL	2.2618847736264014	0
+TYSSCRERL	3.196000128795728	0
+TYSSCRFRF	3.678629266891776	0
+TYSSSEESI	-1.029423306000927	0
+TYSSSGFSL	0.43801710933836974	0