example_sequence_prediction_with_cnn_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "# Goal: Quickly predict the ON and OFF values for sequences in bulk by using trained model. \n",
    "\n",
    "### Instructions: Please change the file_name in the second code block to sequences you are interested in predicting the ON and OFF values for. The format should be a .csv file with one column, and the column should not have a header or column title."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "# import statements \n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from tensorflow.python.saved_model import tag_constants\n",
    "%matplotlib inline\n",
    "\n",
    "import keras as keras\n",
    "from keras.models import load_model\n",
    "\n",
    "from pysster.One_Hot_Encoder import One_Hot_Encoder\n",
    "from sklearn import preprocessing\n",
    "from keras.utils import to_categorical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 1: Load in sequence data. \n",
    "## Change file_name here!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>switch_sequence</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AAAAAAAAAAAAAAAAAATGGAAAACAGTTAACAGAGGAGAAACTG...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AAAAAAAAAAAAATGGAAAACAGTTACTAAAACAGAGGAGATTAGT...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AAAAAAAAATTACTACTATTGTTAATTTAGAACAGAGGAGACTAAA...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     switch_sequence\n",
       "0  AAAAAAAAAAAAAAAAAATGGAAAACAGTTAACAGAGGAGAAACTG...\n",
       "1  AAAAAAAAAAAAATGGAAAACAGTTACTAAAACAGAGGAGATTAGT...\n",
       "2  AAAAAAAAATTACTACTATTGTTAATTTAGAACAGAGGAGACTAAA..."
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# enter a .csv with sequences\n",
    "data_dir = 'data/toehold_sequences/'\n",
    "file_name = 'example_switches.csv' # CHANGE FILENAME!\n",
    "data_df = pd.read_csv(data_dir + file_name,sep=',', header=None)\n",
    "data_df.rename(columns = {0:'switch_sequence'}, inplace = True) \n",
    "data_df.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Toehold length:  59\n",
      "Number of sequences:  99\n"
     ]
    }
   ],
   "source": [
    "seqs = data_df['switch_sequence']\n",
    "seq_len = len(seqs[0])\n",
    "print('Toehold length: ', seq_len)\n",
    "num_seqs = len(data_df)\n",
    "print('Number of sequences: ', num_seqs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 2. Transform Data. One-hot encode sequences and extact target on and off values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create DNA alphabet- may need to change if you have RNA toeholds. Just change to 'AUCG' in the first line\n",
    "alph_letters = sorted('ATCG')\n",
    "alph = list(alph_letters)\n",
    "\n",
    "# one-hot encode with pysster (very fast and simple encoding)  \n",
    "one = One_Hot_Encoder(alph_letters)\n",
    "def _get_one_hot_encoding(seq):\n",
    "    one_hot_seq = one.encode(seq)                         \n",
    "    return one_hot_seq"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 3. Load in final models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_with_given_model(final_model_path, final_weights_path, seqs):\n",
    "\n",
    "    model = load_model(final_model_path)\n",
    "    model.load_weights(final_weights_path)\n",
    "\n",
    "    X = np.stack([_get_one_hot_encoding(s) for s in seqs]).astype(np.float32)\n",
    "    predictions = model.predict(X)\n",
    "    \n",
    "    onoff = predictions\n",
    "    \n",
    "    data_df = pd.DataFrame(columns=['Toehold','Predicted ON/OFF Ratio'])\n",
    "    data_df['Toehold'] = seqs\n",
    "    num_seqs = len(data_df)\n",
    "    data_df['Predicted ON/OFF Ratio'] = np.reshape(onoff, [num_seqs,])\n",
    "    data_df = data_df.sort_values(by='Predicted ON/OFF Ratio', ascending=False)\n",
    "    return data_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:4432: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:148: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:3733: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:203: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:207: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:216: The name tf.is_variable_initialized is deprecated. Please use tf.compat.v1.is_variable_initialized instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:223: The name tf.variables_initializer is deprecated. Please use tf.compat.v1.variables_initializer instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:1033: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n",
      "\n",
      "WARNING:tensorflow:From /anaconda2/envs/clean_toehold_venv/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:1020: The name tf.assign is deprecated. Please use tf.compat.v1.assign instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# load in model trained on Angenent-Mari et al 2020 data\n",
    "model_dir = 'clean_figures/fig4/models/'\n",
    "final_model_path = model_dir + 'onoff_original_model.h5'\n",
    "final_weights_path = model_dir + 'onoff_original_model_weights.h5'\n",
    "original_model_data_df = predict_with_given_model(final_model_path, final_weights_path, seqs)\n",
    "original_model_data_df = original_model_data_df.rename(columns={\"Toehold\": \"seq\", \"Predicted ON/OFF Ratio\": \"original\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load in model trained only on Green et al 2014 data\n",
    "final_model_path = model_dir + 'only_green_trained_model.h5'\n",
    "final_weights_path = model_dir + 'only_green_trained_model_weights.h5'\n",
    "green_model_data_df = predict_with_given_model(final_model_path, final_weights_path, seqs)\n",
    "green_model_data_df = green_model_data_df.rename(columns={\"Toehold\": \"seq\", \"Predicted ON/OFF Ratio\": \"green\"})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load in model trained with transfer learning\n",
    "final_model_path = model_dir + 'freeze_weights_tf_onoff_model.h5'\n",
    "final_weights_path = model_dir + 'freeze_weights_tf_onoff_model_weights.h5'\n",
    "freeze_weights_data_df = predict_with_given_model(final_model_path, final_weights_path, seqs)\n",
    "freeze_weights_data_df = freeze_weights_data_df.rename(columns={\"Toehold\": \"seq\", \"Predicted ON/OFF Ratio\": \"freeze_weights\"})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 4. Put together a dataframe on ON/OFF values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "preds = original_model_data_df\n",
    "preds = preds.merge(right = green_model_data_df, on = 'seq')\n",
    "preds = preds.merge(right = freeze_weights_data_df, on = 'seq')\n",
    "\n",
    "# let's normalize each column by the max value\n",
    "preds['original_pct_of_max'] = preds['original'] / np.max(preds['original'])\n",
    "preds['green_pct_of_max'] = preds['green'] / np.max(preds['green'])\n",
    "preds['freeze_weights_pct_of_max'] = preds['freeze_weights'] / np.max(preds['freeze_weights'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 5. Look at predictions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "rbs = 'AACAGAGGAGA'\n",
    "start_codon = 'ATG'\n",
    "\n",
    "# Make function to generate reverse compliment of the DNA strand\n",
    "def make_rev_complement(string):\n",
    "    new_str = ''\n",
    "    for s in string:\n",
    "        char = ''\n",
    "        if s == 'A':\n",
    "            char = 'T'\n",
    "        elif s == 'T':\n",
    "            char = 'A'\n",
    "        elif s == 'C':\n",
    "            char = 'G'\n",
    "        elif s == 'G':\n",
    "            char = 'C'\n",
    "        else:\n",
    "            print('UH OH! Character not A, T, C, or G')\n",
    "        new_str += char\n",
    "    new_str = new_str[::-1]\n",
    "    return new_str"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "toeholds_30nt = [make_rev_complement(x[0:30]) for x in preds['seq']]\n",
    "preds['30nt_we_are_sensing'] = toeholds_30nt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>seq</th>\n",
       "      <th>original</th>\n",
       "      <th>green</th>\n",
       "      <th>freeze_weights</th>\n",
       "      <th>original_pct_of_max</th>\n",
       "      <th>green_pct_of_max</th>\n",
       "      <th>freeze_weights_pct_of_max</th>\n",
       "      <th>30nt_we_are_sensing</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AAAAAAATTATAATATTCGTATTAGCATCAAACAGAGGAGATGATG...</td>\n",
       "      <td>0.606177</td>\n",
       "      <td>0.122590</td>\n",
       "      <td>0.402334</td>\n",
       "      <td>0.977306</td>\n",
       "      <td>0.396618</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>TGATGCTAATACGAATATTATAATTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>AAAAAAGACGATCAAGAATCCACAATATCAAACAGAGGAGATGATA...</td>\n",
       "      <td>0.564488</td>\n",
       "      <td>0.106070</td>\n",
       "      <td>0.388935</td>\n",
       "      <td>0.910093</td>\n",
       "      <td>0.343173</td>\n",
       "      <td>0.966695</td>\n",
       "      <td>TGATATTGTGGATTCTTGATCGTCTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>AAAAAATAAAACTAGAGATTGAAACGTGTTAACAGAGGAGAAACAC...</td>\n",
       "      <td>0.620253</td>\n",
       "      <td>0.037205</td>\n",
       "      <td>0.367080</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.120370</td>\n",
       "      <td>0.912375</td>\n",
       "      <td>AACACGTTTCAATCTCTAGTTTTATTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>AAAAAATAGCAAGGTATCTATTGAAGATGTAACAGAGGAGAACATC...</td>\n",
       "      <td>0.591277</td>\n",
       "      <td>0.109067</td>\n",
       "      <td>0.364902</td>\n",
       "      <td>0.953283</td>\n",
       "      <td>0.352866</td>\n",
       "      <td>0.906962</td>\n",
       "      <td>ACATCTTCAATAGATACCTTGCTATTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AAAAAACTCTTTGGTCATGTCTTTCTCTTTAACAGAGGAGAAAAGA...</td>\n",
       "      <td>0.589039</td>\n",
       "      <td>0.015357</td>\n",
       "      <td>0.363877</td>\n",
       "      <td>0.949675</td>\n",
       "      <td>0.049684</td>\n",
       "      <td>0.904416</td>\n",
       "      <td>AAAGAGAAAGACATGACCAAAGAGTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>AAAAAAATGACATTATTTAAGAGCATTCTAAACAGAGGAGATAGAA...</td>\n",
       "      <td>0.022622</td>\n",
       "      <td>0.042017</td>\n",
       "      <td>0.015981</td>\n",
       "      <td>0.036472</td>\n",
       "      <td>0.135937</td>\n",
       "      <td>0.039720</td>\n",
       "      <td>TAGAATGCTCTTAAATAATGTCATTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>AAAAAAAATAACGTAGGACTACTACTTGGAAACAGAGGAGATCCAA...</td>\n",
       "      <td>0.061793</td>\n",
       "      <td>0.244613</td>\n",
       "      <td>0.015027</td>\n",
       "      <td>0.099625</td>\n",
       "      <td>0.791403</td>\n",
       "      <td>0.037348</td>\n",
       "      <td>TCCAAGTAGTAGTCCTACGTTATTTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>AAAAAATGTCTAGTGTATCTAGATGAAAATAACAGAGGAGAATTTT...</td>\n",
       "      <td>0.058335</td>\n",
       "      <td>0.013765</td>\n",
       "      <td>0.013889</td>\n",
       "      <td>0.094051</td>\n",
       "      <td>0.044536</td>\n",
       "      <td>0.034521</td>\n",
       "      <td>ATTTTCATCTAGATACACTAGACATTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>AAAAAACTTGTAAATATTAAAACATATGAAAACAGAGGAGATTCAT...</td>\n",
       "      <td>-0.016873</td>\n",
       "      <td>0.057240</td>\n",
       "      <td>0.002763</td>\n",
       "      <td>-0.027204</td>\n",
       "      <td>0.185191</td>\n",
       "      <td>0.006866</td>\n",
       "      <td>TTCATATGTTTTAATATTTACAAGTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>AAAAAATCTTAACTATATAATATTCTAATAAACAGAGGAGATATTA...</td>\n",
       "      <td>0.002365</td>\n",
       "      <td>0.005108</td>\n",
       "      <td>-0.002922</td>\n",
       "      <td>0.003813</td>\n",
       "      <td>0.016525</td>\n",
       "      <td>-0.007261</td>\n",
       "      <td>TATTAGAATATTATATAGTTAAGATTTTTT</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>99 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  seq  original     green  \\\n",
       "1   AAAAAAATTATAATATTCGTATTAGCATCAAACAGAGGAGATGATG...  0.606177  0.122590   \n",
       "4   AAAAAAGACGATCAAGAATCCACAATATCAAACAGAGGAGATGATA...  0.564488  0.106070   \n",
       "0   AAAAAATAAAACTAGAGATTGAAACGTGTTAACAGAGGAGAAACAC...  0.620253  0.037205   \n",
       "2   AAAAAATAGCAAGGTATCTATTGAAGATGTAACAGAGGAGAACATC...  0.591277  0.109067   \n",
       "3   AAAAAACTCTTTGGTCATGTCTTTCTCTTTAACAGAGGAGAAAAGA...  0.589039  0.015357   \n",
       "..                                                ...       ...       ...   \n",
       "96  AAAAAAATGACATTATTTAAGAGCATTCTAAACAGAGGAGATAGAA...  0.022622  0.042017   \n",
       "84  AAAAAAAATAACGTAGGACTACTACTTGGAAACAGAGGAGATCCAA...  0.061793  0.244613   \n",
       "85  AAAAAATGTCTAGTGTATCTAGATGAAAATAACAGAGGAGAATTTT...  0.058335  0.013765   \n",
       "98  AAAAAACTTGTAAATATTAAAACATATGAAAACAGAGGAGATTCAT... -0.016873  0.057240   \n",
       "97  AAAAAATCTTAACTATATAATATTCTAATAAACAGAGGAGATATTA...  0.002365  0.005108   \n",
       "\n",
       "    freeze_weights  original_pct_of_max  green_pct_of_max  \\\n",
       "1         0.402334             0.977306          0.396618   \n",
       "4         0.388935             0.910093          0.343173   \n",
       "0         0.367080             1.000000          0.120370   \n",
       "2         0.364902             0.953283          0.352866   \n",
       "3         0.363877             0.949675          0.049684   \n",
       "..             ...                  ...               ...   \n",
       "96        0.015981             0.036472          0.135937   \n",
       "84        0.015027             0.099625          0.791403   \n",
       "85        0.013889             0.094051          0.044536   \n",
       "98        0.002763            -0.027204          0.185191   \n",
       "97       -0.002922             0.003813          0.016525   \n",
       "\n",
       "    freeze_weights_pct_of_max             30nt_we_are_sensing  \n",
       "1                    1.000000  TGATGCTAATACGAATATTATAATTTTTTT  \n",
       "4                    0.966695  TGATATTGTGGATTCTTGATCGTCTTTTTT  \n",
       "0                    0.912375  AACACGTTTCAATCTCTAGTTTTATTTTTT  \n",
       "2                    0.906962  ACATCTTCAATAGATACCTTGCTATTTTTT  \n",
       "3                    0.904416  AAAGAGAAAGACATGACCAAAGAGTTTTTT  \n",
       "..                        ...                             ...  \n",
       "96                   0.039720  TAGAATGCTCTTAAATAATGTCATTTTTTT  \n",
       "84                   0.037348  TCCAAGTAGTAGTCCTACGTTATTTTTTTT  \n",
       "85                   0.034521  ATTTTCATCTAGATACACTAGACATTTTTT  \n",
       "98                   0.006866  TTCATATGTTTTAATATTTACAAGTTTTTT  \n",
       "97                  -0.007261  TATTAGAATATTATATAGTTAAGATTTTTT  \n",
       "\n",
       "[99 rows x 8 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds = preds.sort_values('freeze_weights', ascending = False)\n",
    "preds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Part 6: Get best and worst sequences."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>seq</th>\n",
       "      <th>original</th>\n",
       "      <th>green</th>\n",
       "      <th>freeze_weights</th>\n",
       "      <th>original_pct_of_max</th>\n",
       "      <th>green_pct_of_max</th>\n",
       "      <th>freeze_weights_pct_of_max</th>\n",
       "      <th>30nt_we_are_sensing</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AAAAAAATTATAATATTCGTATTAGCATCAAACAGAGGAGATGATG...</td>\n",
       "      <td>0.606177</td>\n",
       "      <td>0.122590</td>\n",
       "      <td>0.402334</td>\n",
       "      <td>0.977306</td>\n",
       "      <td>0.396618</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>TGATGCTAATACGAATATTATAATTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>AAAAAATACATGCTCCTTTAATTTTGTTAAAACAGAGGAGATTAAC...</td>\n",
       "      <td>0.531353</td>\n",
       "      <td>0.214239</td>\n",
       "      <td>0.352012</td>\n",
       "      <td>0.856672</td>\n",
       "      <td>0.693135</td>\n",
       "      <td>0.874925</td>\n",
       "      <td>TTAACAAAATTAAAGGAGCATGTATTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>AAAAAACACCTTCTCAACTTTAGATGCTTGAACAGAGGAGACAAGC...</td>\n",
       "      <td>0.390108</td>\n",
       "      <td>0.309088</td>\n",
       "      <td>0.286906</td>\n",
       "      <td>0.628950</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.713103</td>\n",
       "      <td>CAAGCATCTAAAGTTGAGAAGGTGTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>AAAAAAAAAAAAATGGAAAACAGTTACTAAAACAGAGGAGATTAGT...</td>\n",
       "      <td>0.419171</td>\n",
       "      <td>0.121461</td>\n",
       "      <td>0.260160</td>\n",
       "      <td>0.675807</td>\n",
       "      <td>0.392966</td>\n",
       "      <td>0.646627</td>\n",
       "      <td>TTAGTAACTGTTTTCCATTTTTTTTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>AAAAAAGGATTTCTCACCATAGCCAGCATCAACAGAGGAGAGATGC...</td>\n",
       "      <td>0.294863</td>\n",
       "      <td>0.135844</td>\n",
       "      <td>0.158091</td>\n",
       "      <td>0.475392</td>\n",
       "      <td>0.439501</td>\n",
       "      <td>0.392935</td>\n",
       "      <td>GATGCTGGCTATGGTGAGAAATCCTTTTTT</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  seq  original     green  \\\n",
       "1   AAAAAAATTATAATATTCGTATTAGCATCAAACAGAGGAGATGATG...  0.606177  0.122590   \n",
       "7   AAAAAATACATGCTCCTTTAATTTTGTTAAAACAGAGGAGATTAAC...  0.531353  0.214239   \n",
       "25  AAAAAACACCTTCTCAACTTTAGATGCTTGAACAGAGGAGACAAGC...  0.390108  0.309088   \n",
       "20  AAAAAAAAAAAAATGGAAAACAGTTACTAAAACAGAGGAGATTAGT...  0.419171  0.121461   \n",
       "41  AAAAAAGGATTTCTCACCATAGCCAGCATCAACAGAGGAGAGATGC...  0.294863  0.135844   \n",
       "\n",
       "    freeze_weights  original_pct_of_max  green_pct_of_max  \\\n",
       "1         0.402334             0.977306          0.396618   \n",
       "7         0.352012             0.856672          0.693135   \n",
       "25        0.286906             0.628950          1.000000   \n",
       "20        0.260160             0.675807          0.392966   \n",
       "41        0.158091             0.475392          0.439501   \n",
       "\n",
       "    freeze_weights_pct_of_max             30nt_we_are_sensing  \n",
       "1                    1.000000  TGATGCTAATACGAATATTATAATTTTTTT  \n",
       "7                    0.874925  TTAACAAAATTAAAGGAGCATGTATTTTTT  \n",
       "25                   0.713103  CAAGCATCTAAAGTTGAGAAGGTGTTTTTT  \n",
       "20                   0.646627  TTAGTAACTGTTTTCCATTTTTTTTTTTTT  \n",
       "41                   0.392935  GATGCTGGCTATGGTGAGAAATCCTTTTTT  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get good toeholds\n",
    "num_best = 5\n",
    "threshold = 1\n",
    "increment = 0.0001\n",
    "\n",
    "while True:\n",
    "    good_toeholds = preds[preds['original_pct_of_max'] > threshold]\n",
    "    # We note that the Green model is likely enormously underfit but we had an abundance of toeholds pass both the\n",
    "    # original model and transfer learning model threshold so decided to use another filtering step\n",
    "    good_toeholds = good_toeholds[good_toeholds['green_pct_of_max'] > threshold]\n",
    "    good_toeholds = good_toeholds[good_toeholds['freeze_weights_pct_of_max'] > threshold]\n",
    "    if (len(good_toeholds)) >= num_best:\n",
    "        break\n",
    "    else:\n",
    "        threshold = threshold - increment\n",
    "\n",
    "good_toeholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>seq</th>\n",
       "      <th>original</th>\n",
       "      <th>green</th>\n",
       "      <th>freeze_weights</th>\n",
       "      <th>original_pct_of_max</th>\n",
       "      <th>green_pct_of_max</th>\n",
       "      <th>freeze_weights_pct_of_max</th>\n",
       "      <th>30nt_we_are_sensing</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>94</th>\n",
       "      <td>AAAAAATACATATATTAGAATGAATACAACAACAGAGGAGAGTTGT...</td>\n",
       "      <td>0.039368</td>\n",
       "      <td>0.004367</td>\n",
       "      <td>0.035933</td>\n",
       "      <td>0.063471</td>\n",
       "      <td>0.014129</td>\n",
       "      <td>0.089311</td>\n",
       "      <td>GTTGTATTCATTCTAATATATGTATTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>AAAAAAAGAAATTGGCCACAGGACCAAAGGAACAGAGGAGACCTTT...</td>\n",
       "      <td>0.048591</td>\n",
       "      <td>-0.002720</td>\n",
       "      <td>0.028047</td>\n",
       "      <td>0.078340</td>\n",
       "      <td>-0.008800</td>\n",
       "      <td>0.069711</td>\n",
       "      <td>CCTTTGGTCCTGTGGCCAATTTCTTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>AAAAAACTAGTCCATTTAAGAGGTATGTAGAACAGAGGAGACTACA...</td>\n",
       "      <td>0.043079</td>\n",
       "      <td>0.013636</td>\n",
       "      <td>0.016751</td>\n",
       "      <td>0.069454</td>\n",
       "      <td>0.044115</td>\n",
       "      <td>0.041634</td>\n",
       "      <td>CTACATACCTCTTAAATGGACTAGTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>AAAAAAATGACATTATTTAAGAGCATTCTAAACAGAGGAGATAGAA...</td>\n",
       "      <td>0.022622</td>\n",
       "      <td>0.042017</td>\n",
       "      <td>0.015981</td>\n",
       "      <td>0.036472</td>\n",
       "      <td>0.135937</td>\n",
       "      <td>0.039720</td>\n",
       "      <td>TAGAATGCTCTTAAATAATGTCATTTTTTT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>AAAAAATCTTAACTATATAATATTCTAATAAACAGAGGAGATATTA...</td>\n",
       "      <td>0.002365</td>\n",
       "      <td>0.005108</td>\n",
       "      <td>-0.002922</td>\n",
       "      <td>0.003813</td>\n",
       "      <td>0.016525</td>\n",
       "      <td>-0.007261</td>\n",
       "      <td>TATTAGAATATTATATAGTTAAGATTTTTT</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  seq  original     green  \\\n",
       "94  AAAAAATACATATATTAGAATGAATACAACAACAGAGGAGAGTTGT...  0.039368  0.004367   \n",
       "87  AAAAAAAGAAATTGGCCACAGGACCAAAGGAACAGAGGAGACCTTT...  0.048591 -0.002720   \n",
       "91  AAAAAACTAGTCCATTTAAGAGGTATGTAGAACAGAGGAGACTACA...  0.043079  0.013636   \n",
       "96  AAAAAAATGACATTATTTAAGAGCATTCTAAACAGAGGAGATAGAA...  0.022622  0.042017   \n",
       "97  AAAAAATCTTAACTATATAATATTCTAATAAACAGAGGAGATATTA...  0.002365  0.005108   \n",
       "\n",
       "    freeze_weights  original_pct_of_max  green_pct_of_max  \\\n",
       "94        0.035933             0.063471          0.014129   \n",
       "87        0.028047             0.078340         -0.008800   \n",
       "91        0.016751             0.069454          0.044115   \n",
       "96        0.015981             0.036472          0.135937   \n",
       "97       -0.002922             0.003813          0.016525   \n",
       "\n",
       "    freeze_weights_pct_of_max             30nt_we_are_sensing  \n",
       "94                   0.089311  GTTGTATTCATTCTAATATATGTATTTTTT  \n",
       "87                   0.069711  CCTTTGGTCCTGTGGCCAATTTCTTTTTTT  \n",
       "91                   0.041634  CTACATACCTCTTAAATGGACTAGTTTTTT  \n",
       "96                   0.039720  TAGAATGCTCTTAAATAATGTCATTTTTTT  \n",
       "97                  -0.007261  TATTAGAATATTATATAGTTAAGATTTTTT  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get bad toeholds\n",
    "num_worst = 5\n",
    "threshold = -1\n",
    "increment = 0.0001\n",
    "\n",
    "while True:\n",
    "    bad_toeholds = preds[preds['original'] < threshold]\n",
    "    bad_toeholds = bad_toeholds[bad_toeholds['green'] < threshold]\n",
    "    bad_toeholds = bad_toeholds[bad_toeholds['freeze_weights'] < threshold]\n",
    "    if (len(bad_toeholds)) >= num_worst:\n",
    "        break\n",
    "    else:\n",
    "        threshold = threshold + increment\n",
    "\n",
    "bad_toeholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "clean_toehold_venv",
   "language": "python",
   "name": "clean_toehold_venv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}