code/preprocessing_PCA_imputation.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import PowerTransformer\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.impute import KNNImputer"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['ensembl_gene_id', 'loeuf_score', 'ncGERP', 'ncRVIS', 'RVIS_score',\n",
       "       'ncCADD', 'median_tpm', 'num_enh', 'num_super_enh', 'pHaplo', 'pTriplo',\n",
       "       'Episcore', 'pLI', 'tau', 'log_ae', 'log_eqtl'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(r'data\\raw_final_df2.txt', sep='\\t')\n",
    "df = df[df['gene_biotype']=='protein_coding']\n",
    "df1 = df.drop(['gene_biotype','num_enhancers'], axis=1)\n",
    "df1['log_ae'] = np.log(df1['Avg_VG_AE'])\n",
    "df1['log_eqtl'] = np.log(df1['Avg_VG_eQTL'])\n",
    "df1 = df1[df1['Avg_VG_AE'].notna()]\n",
    "df1 = df1.drop(['Avg_VG_AE','Avg_VG_eQTL'], axis=1)\n",
    "df1 = df1.drop_duplicates('ensembl_gene_id')\n",
    "df1.reset_index(drop=True, inplace=True)\n",
    "df1.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# n_components = 10\n",
    "best_df = df1.dropna().reset_index(drop=True)\n",
    "pca = PCA()\n",
    "pca_pipe = Pipeline([('scaler',StandardScaler()),('pca', pca)]) #\n",
    "X_pca = best_df.drop(['ensembl_gene_id','log_ae'], axis=1)\n",
    "y_pca = best_df['log_ae']\n",
    "PC = pca_pipe.fit_transform(X_pca)\n",
    "n_components = len(pca.components_)\n",
    "pca_df = pd.DataFrame(data=PC, columns = ['principal component {}'.format(i+1) for i in range(n_components) ])\n",
    "pca_df['log_ae'] =  y_pca\n",
    "pca_df['ensembl_gene_id'] = best_df['ensembl_gene_id']\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fig. 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "14 14\n"
     ]
    }
   ],
   "source": [
    "cumulative_variance = np.cumsum(pca.explained_variance_ratio_)\n",
    "cumulative_variance = list(cumulative_variance)\n",
    "pcs = ['PC {}'.format(i+1) for i in range(n_components)]\n",
    "plt.figure(figsize = (12,8))\n",
    "plt.bar(pcs, cumulative_variance)\n",
    "plt.xticks(rotation=45, fontsize = 12, ha='right')\n",
    "# plt.xlabel('tissue')\n",
    "plt.ylabel('cumulative explained variance', fontsize=12)\n",
    "plt.grid(axis='y', color='grey', alpha=0.2)\n",
    "# plt.savefig(r'datacollection\\images\\final_report\\PCA_cumulative_variance.png', dpi = 600)\n",
    "# plt.savefig(r'datacollection\\images\\final_report\\PCA_cumulative_variance.svg', dpi = 96, bbox_inches='tight')\n",
    "plt.show()\n",
    "print(len(cumulative_variance), len(pcs))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Venn Diagram:\n",
    "\n",
    "## Fig. 4A\n",
    "* how many values do we have how many can we possibly predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(22895, 16)\n",
      "(12235,)\n",
      "(12807,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#filter for those rows where you have no NA except for NA in log_AE\n",
    "from matplotlib_venn import venn2\n",
    "\n",
    "\n",
    "venn_df = df[df['gene_biotype']=='protein_coding'].drop(['gene_biotype','num_enhancers'], axis=1)\n",
    "venn_df['log_ae'] = np.log(df['Avg_VG_AE'])\n",
    "venn_df['log_eqtl'] = np.log(df['Avg_VG_eQTL'])\n",
    "venn_df = venn_df.drop(['Avg_VG_AE','Avg_VG_eQTL'], axis=1)\n",
    "\n",
    "venn_df = venn_df.drop_duplicates('ensembl_gene_id')\n",
    "venn_df.reset_index(drop=True, inplace=True)\n",
    "print(venn_df.shape)\n",
    "metric_df = venn_df.drop('log_ae',axis=1).dropna()\n",
    "metric_df = metric_df.loc[:,'ensembl_gene_id']\n",
    "print(metric_df.shape)\n",
    "aes = venn_df.loc[:,['ensembl_gene_id','log_ae']].dropna()\n",
    "aes = aes.loc[:,'ensembl_gene_id']\n",
    "print(aes.shape)\n",
    "plt.figure(figsize=(12,8))\n",
    "venn = venn2([set(metric_df), set(aes)],['feature data','AE-VG data'])\n",
    "venn\n",
    "for label in venn.set_labels:\n",
    "    label.set_fontsize(16)\n",
    "for label in venn.subset_labels:\n",
    "    label.set_fontsize(16)\n",
    "# plt.savefig(r'datacollection\\images\\final_report\\avg_venn_before.png', dpi = 300)\n",
    "# plt.savefig(r'datacollection\\images\\final_report\\avg_venn_before.svg', dpi=96, bbox_inches='tight')\n",
    "\n",
    "# print(df.notna().sum(),'\\n', df1.shape)\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# KNN Imputation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def count_na(x):\n",
    "    count = 0 \n",
    "    for i in x: #iterate over the rows of a df. i is a row value (i.e. a field)\n",
    "        \n",
    "        if pd.isna(i):\n",
    "            count += 1 #if a NA is encountered \n",
    "    return count \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   ensembl_gene_id  loeuf_score    ncGERP    ncRVIS  RVIS_score    ncCADD  \\\n",
      "0  ENSG00000102699        0.837 -0.513691  0.090132    4.627272  5.991149   \n",
      "1  ENSG00000027001        1.035 -0.275592 -0.342196    0.624650  6.642385   \n",
      "2  ENSG00000102781        0.439 -0.333999  2.350715   -0.159705  5.639629   \n",
      "3  ENSG00000133101        0.680  0.355963  0.377395   -0.492218  7.921371   \n",
      "4  ENSG00000132952        0.456  0.965914 -0.603996    0.453744  9.363447   \n",
      "\n",
      "   median_tpm  num_enh  num_super_enh    pHaplo   pTriplo  Episcore  \\\n",
      "0     23.8300     59.0            0.0  0.350075  0.332723  0.222067   \n",
      "1      8.7450      0.0            0.0  0.371589  0.589904  0.221000   \n",
      "2      5.9810     91.0            1.0  0.414764  0.715844  0.546933   \n",
      "3      0.6567      0.0            0.0  0.294731  0.301389  0.577117   \n",
      "4     15.6100    136.0            3.0  0.706093  0.555769  0.809633   \n",
      "\n",
      "            pLI       tau  log_eqtl    log_ae    gene_biotype  \n",
      "0  4.820400e-27  0.183037 -6.507604 -5.486193  protein_coding  \n",
      "1  1.587600e-15  0.295105 -5.519982 -3.165733  protein_coding  \n",
      "2  5.316900e-01  0.385531 -6.844555 -6.741216  protein_coding  \n",
      "3  4.311700e-04  0.903231 -4.356312       NaN  protein_coding  \n",
      "4  2.181900e-02  0.277094 -7.097726 -6.295292  protein_coding  \n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv(r'data\\raw_final_df2.txt', sep='\\t')\n",
    "df1 = df[df['gene_biotype']=='protein_coding'].drop(['gene_biotype','num_enhancers'], axis=1) #most of the genetic constraint metrics almost only exist for protein coding genes, so filter the protein_coding genes\n",
    "df1['log_ae'] = np.log(df1['Avg_VG_AE'])\n",
    "df1['log_eqtl'] = np.log(df1['Avg_VG_eQTL'])\n",
    "df1 = df1.drop(['Avg_VG_eQTL','Avg_VG_AE'], axis=1)\n",
    "df1 = df1.drop_duplicates('ensembl_gene_id')\n",
    "df1['na_count'] = df1.apply(lambda x: count_na(x[:]), axis=1)\n",
    "cond = df1.log_eqtl.isna() &  df1.tau.isna() & df1.median_tpm.isna() # boolean for the rows where GTEx v7 data was incomplete and/or the scores have not been calculated  \n",
    "df1 = df1.loc[~cond,:] # exclude those genes where we have no the GTEx data\n",
    "df1 = df1[df1['na_count']<=5] #exclude those genes where we have more than 5 features missing\n",
    "df1.set_index('ensembl_gene_id', inplace=True)\n",
    "\n",
    "\n",
    "X_train = df1.drop(['log_ae','na_count'], axis = 1)\n",
    "features = list(X_train.columns)\n",
    "X_train = np.array(X_train)\n",
    "y = np.array(df1['log_ae'])\n",
    "pipe = Pipeline([('imputer',KNNImputer(n_neighbors=5, weights='uniform'))])#\n",
    "X_imputed_train = pipe.fit_transform(X_train)\n",
    "imputed_train = pd.DataFrame(X_imputed_train, columns = features)\n",
    "imputed_train['log_ae'] = y\n",
    "imputed_train.index = df1.index\n",
    "imputed_train.reset_index(inplace=True) # this makes ensembl_gene_id the first column\n",
    "imputed_train['gene_biotype'] = 'protein_coding'\n",
    "print(imputed_train.head())\n",
    "\n",
    "imputed_train.to_csv(r'data\\imputed_ML_df.csv',sep='\\t',index=False)\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fig. 4B\n",
    "* venn after imputation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(22895, 16)\n",
      "(16957,)\n",
      "(12807,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "imputed_df = pd.read_csv(r'data\\imputed_ML_df.csv', sep='\\t')\n",
    "\n",
    "venn_df = df[df['gene_biotype']=='protein_coding'].drop(['gene_biotype','num_enhancers'], axis=1)\n",
    "venn_df['log_ae'] = np.log(df['Avg_VG_AE'])\n",
    "venn_df['log_eqtl'] = np.log(df['Avg_VG_eQTL'])\n",
    "venn_df = venn_df.drop(['Avg_VG_AE','Avg_VG_eQTL'], axis=1)\n",
    "venn_df = venn_df.drop_duplicates('ensembl_gene_id')\n",
    "venn_df.reset_index(drop=True, inplace=True)\n",
    "print(venn_df.shape)\n",
    "metric_df = imputed_df.drop('log_ae',axis=1).dropna()\n",
    "metric_df = imputed_df.loc[:,'ensembl_gene_id']\n",
    "print(metric_df.shape)\n",
    "aes = venn_df.loc[:,['ensembl_gene_id','log_ae']].dropna()\n",
    "aes = aes.loc[:,'ensembl_gene_id']\n",
    "print(aes.shape)\n",
    "plt.figure(figsize=(12,8))\n",
    "venn = venn2([set(metric_df), set(aes)],['feature data','AE-VG data'])\n",
    "for label in venn.set_labels:\n",
    "    label.set_fontsize(16)\n",
    "for label in venn.subset_labels:\n",
    "    label.set_fontsize(16)\n",
    "venn\n",
    "# plt.savefig(r'datacollection\\images\\final_report\\avg_venn_after.png', dpi = 300)\n",
    "# plt.savefig(r'datacollection\\images\\final_report\\avg_venn_after.svg', dpi=96, bbox_inches='tight')\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.10.4 ('master1')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.4"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "e98864258b4abf79abfb3549cf120b460ab49efdcd183889a03baebd696ce50f"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}