{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Using TensorFlow backend.\n" ] } ], "source": [ "import keras\n", "from keras.models import Sequential, Model, load_model\n", "\n", "import os\n", "import pickle\n", "import numpy as np\n", "import pandas as pd\n", "\n", "import scipy.sparse as sp\n", "import scipy.io as spio\n", "\n", "import matplotlib.pyplot as plt\n", "\n", "from scrambler.models import *\n", "from scrambler.utils import OneHotEncoder, get_sequence_masks\n", "from scrambler.visualizations import plot_protein_logo, plot_protein_importance_scores\n", "\n", "from ppi_utils import load_ppi_data, load_ppi_predictor, animate_ppi_example\n" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Training set size = 40000\n", "Test set size = 4000\n", "x_1_train.shape = (40000, 1, 81, 20)\n", "x_2_train.shape = (40000, 1, 81, 20)\n", "l_1_train.shape = (40000, 1)\n", "l_2_train.shape = (40000, 1)\n", "y_train.shape = (40000, 1)\n", "x_1_test.shape = (4000, 1, 81, 20)\n", "x_2_test.shape = (4000, 1, 81, 20)\n", "l_1_test.shape = (4000, 1)\n", "l_2_test.shape = (4000, 1)\n", "y_test.shape = (4000, 1)\n" ] } ], "source": [ "#Load PPI data and predictor\n", "\n", "seq_length = 81\n", "\n", "residue_map = {'D': 0, 'E': 1, 'V': 2, 'K': 3, 'R': 4, 'L': 5, 'S': 6, 'T': 7, 'N': 8, 'H': 9, 'A': 10, 'I': 11, 'G': 12, 'P': 13, 'Q': 14, 'Y': 15, 'W': 16, 'M': 17, 'F': 18, '#': 19}\n", "\n", "encoder = OneHotEncoder(seq_length, residue_map)\n", "\n", "train_data_path = 'coiled_coil_binders_big_set_train.csv'\n", "test_data_path = 'coiled_coil_binders_big_set_test.csv'\n", "\n", "x_1_train, x_2_train, l_1_train, l_2_train, y_train, x_1_test, x_2_test, l_1_test, l_2_test, y_test = load_ppi_data(train_data_path, test_data_path, encoder)\n", "\n", "predictor_path = 'saved_models/ppi_rnn_baker_big_set_5x_negatives_classifier_symmetric_drop_25_5x_negatives_balanced_partitioned_data_epoch_10.h5'\n", "\n", "predictor = load_ppi_predictor(predictor_path)\n" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "#Define sequence template and background\n", "\n", "#Define sequence templates\n", "\n", "sequence_templates = [\n", " '$' * i + '#' * (seq_length - i)\n", " for i in range(seq_length+1)\n", "]\n", "\n", "onehot_templates = [encoder(sequence_template)[None, ...] for sequence_template in sequence_templates]\n", "\n", "sequence_masks = [\n", " np.array([1 if sequence_templates[i][j] == '$' else 0 for j in range(len(sequence_templates[i]))])\n", " for i in range(seq_length+1)\n", "]\n", "\n", "#Calculate background distributions\n", "\n", "pseudo_count = 0.1\n", "\n", "x_means = []\n", "for i in range(seq_length + 1) :\n", " x_train_len = x_1_train[np.ravel(l_1_train) == i, ...]\n", " \n", " if x_train_len.shape[0] > 0 :\n", " x_mean_len = (np.sum(x_train_len, axis=(0, 1)) + pseudo_count) / (np.sum(x_train_len, axis=(0, 1, 3)).reshape(-1, 1) + 20. * pseudo_count)\n", " x_means.append(x_mean_len)\n", " else :\n", " x_means.append(np.ones((x_1_train.shape[2], x_1_train.shape[3])))\n" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "#Visualize a few background sequence distributions\n", "\n", "visualize_len = 67\n", "\n", "plot_protein_logo(residue_map, np.copy(x_means[visualize_len]), sequence_template=sequence_templates[visualize_len], figsize=(12, 1), logo_height=1.0, plot_start=0, plot_end=81)\n", "\n", "visualize_len = 72\n", "\n", "plot_protein_logo(residue_map, np.copy(x_means[visualize_len]), sequence_template=sequence_templates[visualize_len], figsize=(12, 1), logo_height=1.0, plot_start=0, plot_end=81)\n", "\n", "visualize_len = 81\n", "\n", "plot_protein_logo(residue_map, np.copy(x_means[visualize_len]), sequence_template=sequence_templates[visualize_len], figsize=(12, 1), logo_height=1.0, plot_start=0, plot_end=81)\n" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[Length = 67] Mean KL Div against background (bits) = 1.7578635612470066\n", "[Length = 68] Mean KL Div against background (bits) = 2.5465825890891662\n", "[Length = 69] Mean KL Div against background (bits) = 2.6944880577266086\n", "[Length = 70] Mean KL Div against background (bits) = 2.612709483547898\n", "[Length = 71] Mean KL Div against background (bits) = 2.6518119291022706\n", "[Length = 72] Mean KL Div against background (bits) = 2.68691569539752\n", "[Length = 73] Mean KL Div against background (bits) = 2.660946742737564\n", "[Length = 74] Mean KL Div against background (bits) = 2.732356399534248\n", "[Length = 75] Mean KL Div against background (bits) = 2.6088529559705993\n", "[Length = 76] Mean KL Div against background (bits) = 2.6667669150336644\n", "[Length = 77] Mean KL Div against background (bits) = 2.673211580868109\n", "[Length = 78] Mean KL Div against background (bits) = 2.6430096077067344\n", "[Length = 79] Mean KL Div against background (bits) = 2.634039654091539\n", "[Length = 80] Mean KL Div against background (bits) = 2.6873820395203865\n", "[Length = 81] Mean KL Div against background (bits) = 2.509856290845434\n" ] } ], "source": [ "#Calculate mean training set kl-divergence against background\n", "\n", "mean_kl_divs = []\n", "\n", "for i in range(seq_length + 1) :\n", " x_train_len = x_1_train[np.ravel(l_1_train) == i, ...]\n", " \n", " if x_train_len.shape[0] > 0 :\n", " x_train_clipped_len = np.clip(np.copy(x_train_len[:, 0, :, :]), 1e-8, 1. - 1e-8)\n", "\n", " kl_divs = np.sum(x_train_clipped_len * np.log(x_train_clipped_len / np.tile(np.expand_dims(x_means[i], axis=0), (x_train_clipped_len.shape[0], 1, 1))), axis=-1) / np.log(2.0)\n", "\n", " x_mean_kl_divs = np.sum(kl_divs * sequence_masks[i], axis=-1) / np.sum(sequence_masks[i])\n", " x_mean_kl_div = np.mean(x_mean_kl_divs)\n", "\n", " mean_kl_divs.append(x_mean_kl_div)\n", " \n", " print(\"[Length = \" + str(i) + \"] Mean KL Div against background (bits) = \" + str(x_mean_kl_div))\n", " else :\n", " mean_kl_divs.append(0)\n" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "#Build scrambler\n", "\n", "#Scrambler network configuration\n", "network_config = {\n", " 'n_groups' : 5,\n", " 'n_resblocks_per_group' : 4,\n", " 'n_channels' : 32,\n", " 'window_size' : 3,\n", " 'dilation_rates' : [1, 2, 4, 2, 1],\n", " 'drop_rate' : 0.0,\n", " 'norm_mode' : 'instance',\n", " 'mask_smoothing' : False,\n", " 'mask_smoothing_window_size' : 5,\n", " 'mask_smoothing_std' : 1.,\n", " 'mask_drop_scales' : [1, 5],\n", " 'mask_min_drop_rate' : 0.0,\n", " 'mask_max_drop_rate' : 0.5,\n", " 'label_input' : False\n", "}\n", "\n", "#Initialize scrambler\n", "scrambler = Scrambler(\n", " n_inputs=2,\n", " multi_input_mode='siamese',\n", " scrambler_mode='inclusion',\n", " input_size_x=1,\n", " input_size_y=81,\n", " n_out_channels=20,\n", " input_templates=onehot_templates,\n", " input_backgrounds=x_means,\n", " batch_size=32,\n", " n_samples=32,\n", " sample_mode='gumbel',\n", " zeropad_input=True,\n", " mask_dropout=False,\n", " network_config=network_config\n", ")\n" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "scrolled": true }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Train on 40000 samples, validate on 4000 samples\n", "Epoch 1/10\n", "40000/40000 [==============================] - 359s 9ms/step - loss: 0.7755 - nll_loss: 0.7058 - entropy_loss: 0.0697 - val_loss: 0.6767 - val_nll_loss: 0.6257 - val_entropy_loss: 0.0510\n", "Epoch 2/10\n", "40000/40000 [==============================] - 531s 13ms/step - loss: 0.5968 - nll_loss: 0.5405 - entropy_loss: 0.0563 - val_loss: 0.5641 - val_nll_loss: 0.5184 - val_entropy_loss: 0.0456\n", "Epoch 3/10\n", "40000/40000 [==============================] - 691s 17ms/step - loss: 0.5364 - nll_loss: 0.4888 - entropy_loss: 0.0476 - val_loss: 0.5247 - val_nll_loss: 0.4885 - val_entropy_loss: 0.0363\n", "Epoch 4/10\n", "40000/40000 [==============================] - 699s 17ms/step - loss: 0.5017 - nll_loss: 0.4613 - entropy_loss: 0.0405 - val_loss: 0.5065 - val_nll_loss: 0.4810 - val_entropy_loss: 0.0255\n", "Epoch 5/10\n", "40000/40000 [==============================] - 705s 18ms/step - loss: 0.4806 - nll_loss: 0.4433 - entropy_loss: 0.0373 - val_loss: 0.4785 - val_nll_loss: 0.4479 - val_entropy_loss: 0.0306\n", "Epoch 6/10\n", "40000/40000 [==============================] - 710s 18ms/step - loss: 0.4639 - nll_loss: 0.4287 - entropy_loss: 0.0352 - val_loss: 0.4665 - val_nll_loss: 0.4322 - val_entropy_loss: 0.0342\n", "Epoch 7/10\n", "40000/40000 [==============================] - 698s 17ms/step - loss: 0.4497 - nll_loss: 0.4155 - entropy_loss: 0.0342 - val_loss: 0.4582 - val_nll_loss: 0.4078 - val_entropy_loss: 0.0504\n", "Epoch 8/10\n", "40000/40000 [==============================] - 702s 18ms/step - loss: 0.4383 - nll_loss: 0.4060 - entropy_loss: 0.0323 - val_loss: 0.4458 - val_nll_loss: 0.4117 - val_entropy_loss: 0.0341\n", "Epoch 9/10\n", "40000/40000 [==============================] - 702s 18ms/step - loss: 0.4305 - nll_loss: 0.3984 - entropy_loss: 0.0321 - val_loss: 0.4429 - val_nll_loss: 0.4212 - val_entropy_loss: 0.0218\n", "Epoch 10/10\n", "40000/40000 [==============================] - 697s 17ms/step - loss: 0.4248 - nll_loss: 0.3931 - entropy_loss: 0.0317 - val_loss: 0.4345 - val_nll_loss: 0.4090 - val_entropy_loss: 0.0255\n" ] } ], "source": [ "#Train scrambler\n", "\n", "n_epochs = 10\n", "\n", "train_history = scrambler.train(\n", " predictor,\n", " [x_1_train, x_2_train],\n", " y_train,\n", " [x_1_test, x_2_test],\n", " y_test,\n", " n_epochs,\n", " group_train=[l_1_train, l_2_train],\n", " group_test=[l_1_test, l_2_test],\n", " monitor_test_indices=np.arange(32).tolist(),\n", " monitor_batch_freq_dict={0 : 1, 100 : 5, 1250 : 10},\n", " nll_mode='reconstruction',\n", " predictor_task='classification',\n", " entropy_mode='target',\n", " entropy_bits=0.25,\n", " entropy_weight=10.\n", ")\n" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Saved scrambler model at saved_models/ppi_inclusion_scrambler_bits_025_epochs_10.h5 \n" ] } ], "source": [ "#Save scrambler checkpoint\n", "save_dir = 'saved_models'\n", "\n", "model_name = 'ppi_inclusion_scrambler_bits_025_epochs_10'\n", "\n", "if not os.path.isdir(save_dir):\n", " os.makedirs(save_dir)\n", "\n", "model_path = os.path.join(save_dir, model_name + '.h5')\n", "\n", "scrambler.save_model(model_path)\n", "\n", "#Sub-select train history monitor (one example only) to save space\n", "for t in range(len(train_history['monitor_pwms'])) :\n", " for k in range(len(train_history['monitor_pwms'][t])) :\n", " train_history['monitor_pwms'][t][k] = train_history['monitor_pwms'][t][k][0:1, ...]\n", " train_history['monitor_importance_scores'][t][k] = train_history['monitor_importance_scores'][t][k][0:1, ...]\n", "\n", " train_history['monitor_nll_losses'][t] = train_history['monitor_nll_losses'][t][0:1, ...]\n", " train_history['monitor_entropy_losses'][t] = train_history['monitor_entropy_losses'][t][0:1, ...]\n", "\n", "pickle.dump({'train_history' : train_history}, open(save_dir + '/' + model_name + '_train_history.pickle', 'wb'))\n" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Loaded scrambler model from saved_models/ppi_inclusion_scrambler_bits_025_epochs_10.h5 \n" ] } ], "source": [ "#Load models\n", "save_dir = 'saved_models'\n", "\n", "model_name = 'ppi_inclusion_scrambler_bits_025_epochs_10'\n", "\n", "if not os.path.isdir(save_dir):\n", " os.makedirs(save_dir)\n", "\n", "model_path = os.path.join(save_dir, model_name + '.h5')\n", "\n", "scrambler.load_model(model_path)\n" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "#Plot training statistics\n", "\n", "f, (ax1, ax2) = plt.subplots(1, 2, figsize=(2 * 4, 3))\n", "\n", "n_epochs_actual = len(train_history['nll_loss'])\n", "\n", "ax1.plot(np.arange(1, n_epochs_actual + 1), train_history['nll_loss'], linewidth=3, color='green')\n", "ax1.plot(np.arange(1, n_epochs_actual + 1), train_history['val_nll_loss'], linewidth=3, color='orange')\n", "\n", "plt.sca(ax1)\n", "plt.xlabel(\"Epochs\", fontsize=14)\n", "plt.ylabel(\"NLL\", fontsize=14)\n", "plt.xlim(1, n_epochs_actual)\n", "plt.xticks([1, n_epochs_actual], [1, n_epochs_actual], fontsize=12)\n", "plt.yticks(fontsize=12)\n", "\n", "ax2.plot(np.arange(1, n_epochs_actual + 1), train_history['entropy_loss'], linewidth=3, color='green')\n", "ax2.plot(np.arange(1, n_epochs_actual + 1), train_history['val_entropy_loss'], linewidth=3, color='orange')\n", "\n", "plt.sca(ax2)\n", "plt.xlabel(\"Epochs\", fontsize=14)\n", "plt.ylabel(\"Entropy Loss\", fontsize=14)\n", "plt.xlim(1, n_epochs_actual)\n", "plt.xticks([1, n_epochs_actual], [1, n_epochs_actual], fontsize=12)\n", "plt.yticks(fontsize=12)\n", "\n", "plt.tight_layout()\n", "\n", "plt.show()\n" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "4000/4000 [==============================] - 9s 2ms/step\n" ] } ], "source": [ "#Interpret the test set using the trained scrambler\n", "\n", "[pwm_1_test, pwm_2_test], [sample_1_test, sample_2_test], [importance_scores_1_test, importance_scores_2_test] = scrambler.interpret([x_1_test, x_2_test], group=[l_1_test, l_2_test])\n" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Test pair 0:\n", " - Prediction (original) = 0.85\n", " - Predictions (scrambled) = [0.49, 0.26, 0.06, 0.16, 0.03, 0.41, 0.73, 0.7, 0.82, 0.42, 0.11, 0.4, 0.78, 0.71, 0.77, 0.0, 0.38, 0.53, 0.06, 0.86, 0.08, 0.01, 0.83, 0.71, 0.42, 0.11, 0.66, 0.4, 0.01, 0.01, 0.6, 0.71]\n", "Binder 1:\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Binder 2:\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "#Visualize a few reconstructed sequence patterns\n", "\n", "#Color by ground truth hbnet positions for test pair 0\n", "special_sequence_template_1 = (\"TAEELLEV$KK$DRV$KE$LRV$EEILKVVEVLTRGEVSSEVLKRVLRKLEELTDKLRRV$EE$RRVVEKLN\" + \"#\" * 81)[:81]\n", "special_sequence_template_2 = (\"DLEDLLRRLRRLVDE$RRLVEELERV$RRLEKAVRDNEDERELARL$RE$$DI$DK$DKLAREILEVLKRLLERTE\" + \"#\" * 81)[:81]\n", "\n", "score_quantile = 0.95\n", "\n", "plot_examples = [0]\n", "save_examples = []\n", "\n", "for test_ix in plot_examples :\n", " \n", " pwm_1_curr = pwm_1_test[test_ix:test_ix+1, ...] * sequence_masks[l_1_test[test_ix, 0]][None, None, :, None]\n", " pwm_2_curr = pwm_2_test[test_ix:test_ix+1, ...] * sequence_masks[l_2_test[test_ix, 0]][None, None, :, None]\n", " \n", " importance_scores_1_curr = importance_scores_1_test[test_ix:test_ix+1, ...] * sequence_masks[l_1_test[test_ix, 0]][None, None, :, None]\n", " importance_scores_2_curr = importance_scores_2_test[test_ix:test_ix+1, ...] * sequence_masks[l_2_test[test_ix, 0]][None, None, :, None]\n", " \n", " print(\"Test pair \" + str(test_ix) + \":\")\n", " \n", " y_test_hat_ref = predictor.predict(x=[x_1_test[test_ix:test_ix+1, ...], x_2_test[test_ix:test_ix+1, ...]], batch_size=1)[0, 0]\n", " y_test_hat = predictor.predict(x=[sample_1_test[test_ix, ...], sample_2_test[test_ix, ...]], batch_size=32)[:32, 0].tolist()\n", " \n", " print(\" - Prediction (original) = \" + str(round(y_test_hat_ref, 2))[:4])\n", " print(\" - Predictions (scrambled) = \" + str([float(str(round(y_test_hat[i], 2))[:4]) for i in range(len(y_test_hat))]))\n", " \n", " save_figs = False\n", " if save_examples is not None and test_ix in save_examples :\n", " save_figs = True\n", " \n", " sequence_template_1 = sequence_templates[l_1_test[test_ix, 0]]\n", " sequence_template_2 = sequence_templates[l_2_test[test_ix, 0]]\n", " if special_sequence_template_1 is not None :\n", " sequence_template_1 = special_sequence_template_1\n", " sequence_template_2 = special_sequence_template_2\n", " \n", " q_1 = np.quantile(importance_scores_1_curr[0, 0, :, :], q=score_quantile)\n", " q_2 = np.quantile(importance_scores_2_curr[0, 0, :, :], q=score_quantile)\n", " \n", " seq_1 = encoder.decode(x_1_test[test_ix, 0, :, :])[:l_1_test[test_ix, 0]]\n", " seq_2 = encoder.decode(x_2_test[test_ix, 0, :, :])[:l_2_test[test_ix, 0]]\n", " \n", " print(\"Binder 1:\")\n", "\n", " plot_protein_logo(residue_map, x_1_test[test_ix, 0, :, :], sequence_template=sequence_template_1.replace('#', '@'), color_reference=['red'], sequence_colors=np.zeros(81, dtype=np.int).tolist(), figsize=(12, 1), plot_start=0, plot_end=81, save_figs=save_figs, fig_name=model_name + \"_test_ix_\" + str(test_ix) + \"_orig_sequence_binder_1\")\n", " plot_protein_logo(residue_map, pwm_1_curr[0, 0, :, :], sequence_template=sequence_template_1.replace('#', '@'), color_reference=['red'], sequence_colors=np.zeros(81, dtype=np.int).tolist(), figsize=(12, 1), plot_start=0, plot_end=81, save_figs=save_figs, fig_name=model_name + \"_test_ix_\" + str(test_ix) + \"_scrambld_pwm_binder_1\")\n", " plot_protein_importance_scores(importance_scores_1_curr[0, 0, :, :].T, seq_1, figsize=(12, 1), score_clip=q_1, sequence_template=sequence_template_1, single_color='red', fixed_sequence_template_scores=False, plot_start=0, plot_end=81, save_figs=save_figs, fig_name=model_name + \"_test_ix_\" + str(test_ix) + \"_scores_binder_1\")\n", "\n", " print(\"Binder 2:\")\n", "\n", " plot_protein_logo(residue_map, x_2_test[test_ix, 0, :, :], sequence_template=sequence_template_2.replace('#', '@'), color_reference=['red'], sequence_colors=np.zeros(81, dtype=np.int).tolist(), figsize=(12, 1), plot_start=0, plot_end=81, save_figs=save_figs, fig_name=model_name + \"_test_ix_\" + str(test_ix) + \"_orig_sequence_binder_2\")\n", " plot_protein_logo(residue_map, pwm_2_curr[0, 0, :, :], sequence_template=sequence_template_2.replace('#', '@'), color_reference=['red'], sequence_colors=np.zeros(81, dtype=np.int).tolist(), figsize=(12, 1), plot_start=0, plot_end=81, save_figs=save_figs, fig_name=model_name + \"_test_ix_\" + str(test_ix) + \"_scrambld_pwm_binder_2\")\n", " plot_protein_importance_scores(importance_scores_2_curr[0, 0, :, :].T, seq_2, figsize=(12, 1), score_clip=q_2, sequence_template=sequence_template_2, single_color='red', fixed_sequence_template_scores=False, plot_start=0, plot_end=81, save_figs=save_figs, fig_name=model_name + \"_test_ix_\" + str(test_ix) + \"_scores_binder_2\")\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Environment (conda_tensorflow_p36)", "language": "python", "name": "conda_tensorflow_p36" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.6.5" } }, "nbformat": 4, "nbformat_minor": 2 }