examples/graph_neural_networks.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import torch\n",
    "from torch_geometric.nn.models import GIN\n",
    "from gpytorch.priors import NormalPrior\n",
    "from gpytorch.kernels import RBFKernel\n",
    "from torch_geometric.nn import global_mean_pool\n",
    "\n",
    "from mobius import Planner, SequenceGA\n",
    "from mobius import Graph, GPGNNModel, ExpectedImprovement\n",
    "from mobius import LinearPeptideEmulator\n",
    "from mobius import homolog_scanning, alanine_scanning\n",
    "from mobius import convert_FASTA_to_HELM"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Initialize MHC emulator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "pssm_files = ['../data/IEDB_MHC_I-2.9_matx_smm_smmpmbec/smmpmbec_matrix/HLA-A-02:01-8.txt',\n",
    "              '../data/IEDB_MHC_I-2.9_matx_smm_smmpmbec/smmpmbec_matrix/HLA-A-02:01-9.txt',\n",
    "              '../data/IEDB_MHC_I-2.9_matx_smm_smmpmbec/smmpmbec_matrix/HLA-A-02:01-10.txt',\n",
    "              '../data/IEDB_MHC_I-2.9_matx_smm_smmpmbec/smmpmbec_matrix/HLA-A-02:01-11.txt']\n",
    "lpe = LinearPeptideEmulator(pssm_files)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define lead peptide and create initial peptide batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "lead_peptide = convert_FASTA_to_HELM('HMTEVVRRC')[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reach max. number of peptides allowed.\n"
     ]
    }
   ],
   "source": [
    "seed_library = [lead_peptide]\n",
    "\n",
    "for seq in alanine_scanning(lead_peptide):\n",
    "    seed_library.append(seq)\n",
    "\n",
    "for seq in homolog_scanning(lead_peptide):\n",
    "    seed_library.append(seq)\n",
    "\n",
    "    if len(seed_library) >= 96:\n",
    "        print('Reach max. number of peptides allowed.')\n",
    "        break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "pic50_seed_library = lpe.score(seed_library)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define Gaussian Process, acquisition function and GA optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "class GNNModel(torch.nn.Module):\n",
    "    def __init__(self, in_channels, hidden_channels, out_channels, num_layers, dropout):\n",
    "        super(GNNModel, self).__init__()\n",
    "\n",
    "        self.model = GIN(in_channels=in_channels, hidden_channels=hidden_channels, num_layers=num_layers, out_channels=out_channels, dropout=dropout)\n",
    "        \n",
    "        self._reset_parameters()\n",
    "\n",
    "    def _reset_parameters(self):\n",
    "        r\"\"\"Initiate parameters in the transformer model.\"\"\"\n",
    "\n",
    "        for p in self.parameters():\n",
    "            if p.dim() > 1:\n",
    "                torch.nn.init.xavier_uniform_(p)\n",
    "\n",
    "    def forward(self, x):\n",
    "        node_attr, edge_index, edge_attr, batch = x.node_attr, x.edge_index, x.edge_attr, x.batch\n",
    "        x = self.model(x=node_attr, edge_index=edge_index, edge_attr=edge_attr, batch=batch)\n",
    "        x = global_mean_pool(x, batch)\n",
    "\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "graph_convertor = Graph(input_type='helm', output_type='pyg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "graphs = graph_convertor.transform(lead_peptide)\n",
    "gnn = GNNModel(in_channels=graphs.node_attr.shape[1], hidden_channels=256, out_channels=256, num_layers=4, dropout=0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "gpmodel = GPGNNModel(kernel=RBFKernel(), model=gnn, transform=graph_convertor, noise_prior=NormalPrior(loc=0, scale=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "yaml_content = \"\"\"\n",
    "design:\n",
    "  monomers:\n",
    "    default: [A, C, D, E, F, G, H, I, K, L, M, N, P, Q, R, S, T, V, W, Y]\n",
    "    APOLAR: [A, F, G, I, L, P, V, W]\n",
    "    POLAR: [C, D, E, H, K, N, Q, R, K, S, T, M]\n",
    "    AROMATIC: [F, H, W, Y]\n",
    "    POS_CHARGED: [K, R]\n",
    "    NEG_CHARGED: [D, E]\n",
    "  polymers:\n",
    "    - PEPTIDE1{X.M.X.X.X.X.X.X.X}$$$$V2.0:\n",
    "        PEPTIDE1:\n",
    "          1: [AROMATIC, NEG_CHARGED]\n",
    "          4: POLAR\n",
    "          9: [A, V, I, L, M, T]\n",
    "filters:\n",
    "  - class_path: mobius.PeptideSelfAggregationFilter\n",
    "  - class_path: mobius.PeptideSolubilityFilter\n",
    "    init_args:\n",
    "      hydrophobe_ratio: 0.5\n",
    "      charged_per_amino_acids: 5\n",
    "\"\"\"\n",
    "\n",
    "with open('design_protocol.yaml', 'w') as f:\n",
    "    f.write(yaml_content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "ei = ExpectedImprovement(gpmodel, maximize=False)\n",
    "optimizer = SequenceGA(algorithm='GA', period=15, n_pop=500, design_protocol_filename='design_protocol.yaml')\n",
    "ps = Planner(ei, optimizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run optimization!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Fitting GPGNN model (cpu): 10step [00:05,  1.96step/s, loss=1.06]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==========================================================================================================================================\n",
      "n_gen  |  n_eval  |     cv_min    |     cv_avg    |        f_avg         |        f_min         |              best sequence              \n",
      "==========================================================================================================================================\n",
      "     1 |        0 |  0.000000E+00 |  0.1428571429 |  8.63289030303030280 |  7.16085999999999956 |      PEPTIDE1{H.M.T.E.V.V.R.R.V}$$$$V2.0\n",
      "     2 |      500 |  0.000000E+00 |  0.0380000000 |  1.18437633877007320 | -5.1026875012989E-03 |      PEPTIDE1{E.M.Y.E.R.S.R.R.A}$$$$V2.0\n",
      "     3 |     1000 |  0.000000E+00 |  0.000000E+00 | -5.4047115435575E-04 | -8.1498524253159E-03 |      PEPTIDE1{E.M.G.K.A.V.G.R.A}$$$$V2.0\n",
      "     4 |     1500 |  0.000000E+00 |  0.000000E+00 | -1.3956728018870E-03 | -1.4248674470108E-02 |      PEPTIDE1{E.M.L.E.K.Q.Q.R.A}$$$$V2.0\n",
      "     5 |     2000 |  0.000000E+00 |  0.000000E+00 | -3.2925416643967E-03 | -3.0138723841268E-02 |      PEPTIDE1{E.M.R.E.G.K.K.R.L}$$$$V2.0\n",
      "     6 |     2500 |  0.000000E+00 |  0.000000E+00 | -6.5501399467074E-03 | -3.7023652139719E-02 |      PEPTIDE1{E.M.R.K.R.K.K.R.L}$$$$V2.0\n",
      "     7 |     3000 |  0.000000E+00 |  0.000000E+00 | -1.1685575322890E-02 | -3.7023652139719E-02 |      PEPTIDE1{E.M.R.K.R.K.K.R.L}$$$$V2.0\n",
      "     8 |     3500 |  0.000000E+00 |  0.000000E+00 | -1.7974995142627E-02 | -4.1291889618074E-02 |      PEPTIDE1{E.M.K.K.R.K.R.R.A}$$$$V2.0\n",
      "     9 |     4000 |  0.000000E+00 |  0.000000E+00 | -2.4644281109335E-02 | -4.2300164237378E-02 |      PEPTIDE1{E.M.K.K.R.K.K.R.A}$$$$V2.0\n",
      "    10 |     4500 |  0.000000E+00 |  0.000000E+00 | -2.9672335602168E-02 | -4.2300164237378E-02 |      PEPTIDE1{E.M.K.K.R.K.K.R.A}$$$$V2.0\n",
      "    11 |     5000 |  0.000000E+00 |  0.000000E+00 | -3.2907093978999E-02 | -4.2300273474742E-02 |      PEPTIDE1{E.M.K.K.R.K.R.K.A}$$$$V2.0\n",
      "    12 |     5500 |  0.000000E+00 |  0.000000E+00 | -3.5271455118714E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    13 |     6000 |  0.000000E+00 |  0.000000E+00 | -3.6771469459523E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    14 |     6500 |  0.000000E+00 |  0.000000E+00 | -3.7444195183166E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    15 |     7000 |  0.000000E+00 |  0.000000E+00 | -3.7610070066623E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    16 |     7500 |  0.000000E+00 |  0.000000E+00 | -3.7635554599714E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    17 |     8000 |  0.000000E+00 |  0.000000E+00 | -3.7658759371174E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    18 |     8500 |  0.000000E+00 |  0.000000E+00 | -3.7676657624938E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    19 |     9000 |  0.000000E+00 |  0.000000E+00 | -3.7695498325620E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    20 |     9500 |  0.000000E+00 |  0.000000E+00 | -3.7703468554070E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    21 |    10000 |  0.000000E+00 |  0.000000E+00 | -3.7709459789924E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    22 |    10500 |  0.000000E+00 |  0.000000E+00 | -3.7715369483709E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    23 |    11000 |  0.000000E+00 |  0.000000E+00 | -3.7718686001850E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    24 |    11500 |  0.000000E+00 |  0.000000E+00 | -3.7725664203873E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    25 |    12000 |  0.000000E+00 |  0.000000E+00 | -3.7725665289588E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    26 |    12500 |  0.000000E+00 |  0.000000E+00 | -3.7725665999216E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "    27 |    13000 |  0.000000E+00 |  0.000000E+00 | -3.7728612671482E-02 | -4.3476418760304E-02 |      PEPTIDE1{E.M.K.K.R.K.K.K.A}$$$$V2.0\n",
      "Best peptide found so far: PEPTIDE1{H.M.T.E.V.V.R.R.V}$$$$V2.0 / 2.171\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Fitting GPGNN model (cpu): 9step [00:09,  1.10s/step, loss=0.605]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==========================================================================================================================================\n",
      "n_gen  |  n_eval  |     cv_min    |     cv_avg    |        f_avg         |        f_min         |              best sequence              \n",
      "==========================================================================================================================================\n",
      "     1 |        0 |  0.000000E+00 |  0.0581395349 |  1.0933032839506E+01 |  8.25885999999999854 |      PEPTIDE1{H.M.T.E.V.V.R.R.V}$$$$V2.0\n",
      "     2 |      500 |  0.000000E+00 |  0.000000E+00 |  1.44338389508009080 | -1.9727267765940E-01 |      PEPTIDE1{E.M.W.H.L.G.N.K.A}$$$$V2.0\n",
      "     3 |     1000 |  0.000000E+00 |  0.000000E+00 | -4.4282511959982E-03 | -1.9727267765940E-01 |      PEPTIDE1{E.M.W.H.L.G.N.K.A}$$$$V2.0\n",
      "     4 |     1500 |  0.000000E+00 |  0.000000E+00 | -9.5807475346516E-03 | -1.9727267765940E-01 |      PEPTIDE1{E.M.W.H.L.G.N.K.A}$$$$V2.0\n",
      "     5 |     2000 |  0.000000E+00 |  0.000000E+00 | -1.9295083549351E-02 | -1.9727267765940E-01 |      PEPTIDE1{E.M.W.H.L.G.N.K.A}$$$$V2.0\n",
      "     6 |     2500 |  0.000000E+00 |  0.000000E+00 | -3.8124692128507E-02 | -2.3762680280816E-01 |      PEPTIDE1{E.M.W.D.L.G.H.R.A}$$$$V2.0\n",
      "     7 |     3000 |  0.000000E+00 |  0.000000E+00 | -6.8347587835101E-02 | -2.7912977020150E-01 |      PEPTIDE1{H.M.G.D.L.G.S.R.L}$$$$V2.0\n",
      "     8 |     3500 |  0.000000E+00 |  0.000000E+00 | -1.1183841719607E-01 | -3.3088782704817E-01 |      PEPTIDE1{H.M.G.H.E.S.D.N.A}$$$$V2.0\n",
      "     9 |     4000 |  0.000000E+00 |  0.000000E+00 | -1.5530216470409E-01 | -4.1167569827767E-01 |      PEPTIDE1{H.M.G.E.N.G.N.D.A}$$$$V2.0\n",
      "    10 |     4500 |  0.000000E+00 |  0.000000E+00 | -2.0020888381975E-01 | -4.6616045021892E-01 |      PEPTIDE1{E.M.H.D.D.G.N.D.L}$$$$V2.0\n",
      "    11 |     5000 |  0.000000E+00 |  0.000000E+00 | -2.4264201531435E-01 | -4.6616045021892E-01 |      PEPTIDE1{E.M.H.D.D.G.N.D.L}$$$$V2.0\n",
      "    12 |     5500 |  0.000000E+00 |  0.000000E+00 | -2.8242246096850E-01 | -4.6616045021892E-01 |      PEPTIDE1{E.M.H.D.D.G.N.D.L}$$$$V2.0\n",
      "    13 |     6000 |  0.000000E+00 |  0.000000E+00 | -3.2137979438008E-01 | -4.7712264815574E-01 |      PEPTIDE1{E.M.G.D.W.G.D.N.L}$$$$V2.0\n",
      "    14 |     6500 |  0.000000E+00 |  0.000000E+00 | -3.5920254261590E-01 | -4.9649926460172E-01 |      PEPTIDE1{E.M.F.D.D.G.F.G.L}$$$$V2.0\n",
      "    15 |     7000 |  0.000000E+00 |  0.000000E+00 | -3.9542640010409E-01 | -4.9649926460172E-01 |      PEPTIDE1{E.M.F.D.D.G.F.G.L}$$$$V2.0\n",
      "    16 |     7500 |  0.000000E+00 |  0.000000E+00 | -4.2372650741033E-01 | -4.9649926460172E-01 |      PEPTIDE1{E.M.F.D.D.G.F.G.L}$$$$V2.0\n",
      "    17 |     8000 |  0.000000E+00 |  0.000000E+00 | -4.3929085932755E-01 | -4.9649926460172E-01 |      PEPTIDE1{E.M.F.D.D.G.F.G.L}$$$$V2.0\n",
      "    18 |     8500 |  0.000000E+00 |  0.000000E+00 | -4.5126706543555E-01 | -5.0075735691888E-01 |      PEPTIDE1{E.M.F.H.D.G.F.G.L}$$$$V2.0\n",
      "    19 |     9000 |  0.000000E+00 |  0.000000E+00 | -4.5853183826187E-01 | -5.0075735691888E-01 |      PEPTIDE1{E.M.F.H.D.G.F.G.L}$$$$V2.0\n",
      "    20 |     9500 |  0.000000E+00 |  0.000000E+00 | -4.6466638512204E-01 | -5.0180966813619E-01 |      PEPTIDE1{E.M.F.D.G.F.F.G.L}$$$$V2.0\n",
      "    21 |    10000 |  0.000000E+00 |  0.000000E+00 | -4.6858542870173E-01 | -5.0180966813619E-01 |      PEPTIDE1{E.M.F.D.G.F.F.G.L}$$$$V2.0\n",
      "    22 |    10500 |  0.000000E+00 |  0.000000E+00 | -4.7236463242782E-01 | -5.0180966813619E-01 |      PEPTIDE1{E.M.F.D.G.F.F.G.L}$$$$V2.0\n",
      "    23 |    11000 |  0.000000E+00 |  0.000000E+00 | -4.7505233227166E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    24 |    11500 |  0.000000E+00 |  0.000000E+00 | -4.7743999488324E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    25 |    12000 |  0.000000E+00 |  0.000000E+00 | -4.7910635915290E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    26 |    12500 |  0.000000E+00 |  0.000000E+00 | -4.8019218816693E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    27 |    13000 |  0.000000E+00 |  0.000000E+00 | -4.8086622027782E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    28 |    13500 |  0.000000E+00 |  0.000000E+00 | -4.8158803034837E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    29 |    14000 |  0.000000E+00 |  0.000000E+00 | -4.8215661042514E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    30 |    14500 |  0.000000E+00 |  0.000000E+00 | -4.8262536890940E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    31 |    15000 |  0.000000E+00 |  0.000000E+00 | -4.8314046073813E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    32 |    15500 |  0.000000E+00 |  0.000000E+00 | -4.8343799706991E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    33 |    16000 |  0.000000E+00 |  0.000000E+00 | -4.8384519156605E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    34 |    16500 |  0.000000E+00 |  0.000000E+00 | -4.8432169036329E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    35 |    17000 |  0.000000E+00 |  0.000000E+00 | -4.8459827734257E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    36 |    17500 |  0.000000E+00 |  0.000000E+00 | -4.8484495255896E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    37 |    18000 |  0.000000E+00 |  0.000000E+00 | -4.8500104114093E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "    38 |    18500 |  0.000000E+00 |  0.000000E+00 | -4.8529082961375E-01 | -5.0181290177173E-01 |      PEPTIDE1{E.M.F.D.F.G.F.G.L}$$$$V2.0\n",
      "Best peptide found so far: PEPTIDE1{E.M.F.D.G.F.F.G.L}$$$$V2.0 / 1.554\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Fitting GPGNN model (cpu): 8step [00:15,  1.96s/step, loss=0.819]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==========================================================================================================================================\n",
      "n_gen  |  n_eval  |     cv_min    |     cv_avg    |        f_avg         |        f_min         |              best sequence              \n",
      "==========================================================================================================================================\n",
      "     1 |        0 |  0.000000E+00 |  0.0334572491 |  1.0223271538462E+01 |  7.64185999999999943 |      PEPTIDE1{E.M.F.D.G.F.F.G.L}$$$$V2.0\n",
      "     2 |      500 |  0.000000E+00 |  0.000000E+00 |  1.94253752173670202 | -1.4933464567168E-04 |      PEPTIDE1{E.M.G.D.G.F.G.F.L}$$$$V2.0\n",
      "     3 |     1000 |  0.000000E+00 |  0.000000E+00 | -2.0447516335111E-05 | -1.4933464567168E-04 |      PEPTIDE1{E.M.G.D.G.F.G.F.L}$$$$V2.0\n",
      "     4 |     1500 |  0.000000E+00 |  0.000000E+00 | -3.9507350333109E-05 | -1.5432396002957E-04 |      PEPTIDE1{E.M.F.D.G.F.G.F.L}$$$$V2.0\n",
      "     5 |     2000 |  0.000000E+00 |  0.000000E+00 | -5.9548819200100E-05 | -1.5432396002957E-04 |      PEPTIDE1{E.M.F.D.G.F.G.F.L}$$$$V2.0\n",
      "     6 |     2500 |  0.000000E+00 |  0.000000E+00 | -7.3641465912250E-05 | -2.2491751916057E-04 |      PEPTIDE1{E.M.G.D.G.G.F.G.I}$$$$V2.0\n",
      "     7 |     3000 |  0.000000E+00 |  0.000000E+00 | -8.3674993670976E-05 | -2.2491751916057E-04 |      PEPTIDE1{E.M.G.D.G.G.F.G.I}$$$$V2.0\n",
      "     8 |     3500 |  0.000000E+00 |  0.000000E+00 | -9.0412461534487E-05 | -2.2491751916057E-04 |      PEPTIDE1{E.M.G.D.G.G.F.G.I}$$$$V2.0\n",
      "     9 |     4000 |  0.000000E+00 |  0.000000E+00 | -9.5114256795510E-05 | -2.2491751916057E-04 |      PEPTIDE1{E.M.G.D.G.G.F.G.I}$$$$V2.0\n",
      "    10 |     4500 |  0.000000E+00 |  0.000000E+00 | -9.9059826238334E-05 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    11 |     5000 |  0.000000E+00 |  0.000000E+00 | -1.0179042578872E-04 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    12 |     5500 |  0.000000E+00 |  0.000000E+00 | -1.0469767616884E-04 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    13 |     6000 |  0.000000E+00 |  0.000000E+00 | -1.0689287626597E-04 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    14 |     6500 |  0.000000E+00 |  0.000000E+00 | -1.0850532646612E-04 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    15 |     7000 |  0.000000E+00 |  0.000000E+00 | -1.0928086173178E-04 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    16 |     7500 |  0.000000E+00 |  0.000000E+00 | -1.1044285057115E-04 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    17 |     8000 |  0.000000E+00 |  0.000000E+00 | -1.1123652453140E-04 | -2.2492229454719E-04 |      PEPTIDE1{E.M.G.D.G.F.G.G.I}$$$$V2.0\n",
      "    18 |     8500 |  0.000000E+00 |  0.000000E+00 | -1.1211866502837E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    19 |     9000 |  0.000000E+00 |  0.000000E+00 | -1.1293584480046E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    20 |     9500 |  0.000000E+00 |  0.000000E+00 | -1.1435054545660E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    21 |    10000 |  0.000000E+00 |  0.000000E+00 | -1.1492447872026E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    22 |    10500 |  0.000000E+00 |  0.000000E+00 | -1.1633479977741E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    23 |    11000 |  0.000000E+00 |  0.000000E+00 | -1.1760951635964E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    24 |    11500 |  0.000000E+00 |  0.000000E+00 | -1.1858471104040E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    25 |    12000 |  0.000000E+00 |  0.000000E+00 | -1.1988243248073E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    26 |    12500 |  0.000000E+00 |  0.000000E+00 | -1.2095389754511E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    27 |    13000 |  0.000000E+00 |  0.000000E+00 | -1.2207202139440E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    28 |    13500 |  0.000000E+00 |  0.000000E+00 | -1.2258919045342E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    29 |    14000 |  0.000000E+00 |  0.000000E+00 | -1.2331489516325E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    30 |    14500 |  0.000000E+00 |  0.000000E+00 | -1.2378120960228E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    31 |    15000 |  0.000000E+00 |  0.000000E+00 | -1.2416767244485E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    32 |    15500 |  0.000000E+00 |  0.000000E+00 | -1.2464058943070E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "    33 |    16000 |  0.000000E+00 |  0.000000E+00 | -1.2525320083282E-04 | -2.2492835713956E-04 |      PEPTIDE1{F.M.G.D.G.E.G.G.I}$$$$V2.0\n",
      "Best peptide found so far: PEPTIDE1{F.M.D.E.G.F.F.G.L}$$$$V2.0 / 0.247\n",
      "\n"
     ]
    }
   ],
   "source": [
    "peptides = seed_library.copy()\n",
    "pic50_scores = pic50_seed_library.copy()\n",
    "\n",
    "# Store data for later analysis\n",
    "data = [(0, p, s) for p, s in zip(peptides, pic50_scores)]\n",
    "\n",
    "for i in range(3):\n",
    "    suggested_peptides, _ = ps.recommend(peptides, pic50_scores.reshape(-1, 1), batch_size=96)\n",
    "\n",
    "    # Here you can add whatever methods you want to further filter out peptides\n",
    "\n",
    "    # Virtually test the suggested peptides using the MHC emulator\n",
    "    # You know the drill now, this is for benchmarking or demonstration\n",
    "    # purposes only and should be replaced with actual lab experiments.\n",
    "    pic50_suggested_peptides = lpe.score(suggested_peptides)\n",
    "\n",
    "    peptides = np.concatenate([peptides, suggested_peptides])\n",
    "    pic50_scores = np.concatenate((pic50_scores, pic50_suggested_peptides), axis=0)\n",
    "    data.extend([(i + 1, p, s) for p, s in zip(suggested_peptides, pic50_suggested_peptides)])\n",
    "\n",
    "    best_seq = peptides[np.argmin(pic50_scores)]\n",
    "    best_pic50 = np.min(pic50_scores)\n",
    "    print('Best peptide found so far: %s / %.3f' % (best_seq, best_pic50))\n",
    "    print('')\n",
    "\n",
    "df = pd.DataFrame(data=data, columns=('iter', 'polymer', 'exp_value'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x750 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(20, 7.5))\n",
    "\n",
    "# pandas dataframe (df) columns: iter, peptide, exp_value\n",
    "df['iter'] = df['iter'].replace({0: 'Init.', 1: '1', 2: '2', 3: '3'})\n",
    "\n",
    "sns.lineplot(x='iter', y='exp_value', data=df, ax=ax, errorbar=('ci', 95), err_style='bars', linewidth=3.5, err_kws={'lw': 3.5})\n",
    "\n",
    "ax.set_xlim([-0.9, 3.5])\n",
    "ax.set_ylim([-2, 6.5])\n",
    "ax.set_ylabel('pIC50', fontsize=30)\n",
    "ax.set_xlabel('Generations', fontsize=30)\n",
    "ax.xaxis.set_tick_params(labelsize=30)\n",
    "ax.yaxis.set_tick_params(labelsize=30)\n",
    "\n",
    "sns.despine()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:mambaforge-mobius-ubuntu]",
   "language": "python",
   "name": "conda-env-mambaforge-mobius-ubuntu-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": "3.11.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}