40_multiclass.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Use a one v. rest approach to create a multiclass classifier from LASSO\n",
    "Also try a dedicated multiclass classfier\n",
    "\n",
    "Shaina Lu\n",
    "Zador & Gillis Labs\n",
    "April 2021\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import division\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import scipy as sp\n",
    "from scipy import stats\n",
    "from sklearn import metrics\n",
    "from sklearn.linear_model import Lasso\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.model_selection import train_test_split #stratify train/test split\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def set_style():\n",
    "    #function for matplotlib formatting\n",
    "    plt.style.use(['seaborn-white','seaborn-notebook'])\n",
    "    plt.rcParams['figure.figsize'] = [6,4]\n",
    "    plt.rcParams['axes.spines.top'] = False       #remove top line\n",
    "    plt.rcParams['axes.spines.right'] = False     #remove right line\n",
    "    plt.rcParams['axes.linewidth'] = 2.0          #set weight of axes\n",
    "    plt.rcParams['axes.titlesize'] = 20           #set font size of title\n",
    "    plt.rcParams['axes.labelsize'] = 18           #set font size of x,y labels\n",
    "    plt.rcParams['axes.labelpad'] = 14            #space between labels and axes\n",
    "    plt.rcParams['xtick.labelsize'] = 14          #set x label size\n",
    "    plt.rcParams['ytick.labelsize'] = 14          #set y label size\n",
    "    plt.rcParams['legend.fontsize'] = 16          #set legend font size\n",
    "    \n",
    "set_style()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Read in Data and pre-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#file paths\n",
    "ALLEN_FILT_PATH = \"/home/slu/spatial/data/ABAISH_filt_v6_avgdup.h5\"\n",
    "ONTOLOGY_PATH = \"/data/slu/allen_adult_mouse_ISH/ontologyABA.csv\"\n",
    "ST_CANTIN_FILT_PATH = \"/home/slu/spatial/data/cantin_ST_filt_v2.h5\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "jupyter": {
     "source_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "def read_ABAdata():\n",
    "    \"\"\"read in all ABA datasets needed using pandas\"\"\"\n",
    "    metabrain = pd.read_hdf(ALLEN_FILT_PATH, key='metabrain', mode='r')\n",
    "    voxbrain = pd.read_hdf(ALLEN_FILT_PATH, key='avgvoxbrain', mode='r')\n",
    "    propontvox = pd.read_hdf(ALLEN_FILT_PATH, key='propontology', mode='r')\n",
    "    #geneIDName = pd.read_hdf(ALLEN_FILT_PATH, key='geneIDName', mode='r')\t\n",
    "\n",
    "    return metabrain, voxbrain, propontvox\n",
    "\n",
    "#ST\n",
    "def read_STdata():\n",
    "    \"\"\"read in all ST datasets needed using pandas\"\"\"\n",
    "    STspotsmeta = pd.read_hdf(ST_CANTIN_FILT_PATH, key='STspotsmeta', mode='r')\n",
    "    STspots = pd.read_hdf(ST_CANTIN_FILT_PATH, key='STspots', mode='r')\n",
    "    STpropont = pd.read_hdf(ST_CANTIN_FILT_PATH, key='STpropont', mode='r')\n",
    "    \n",
    "    return STspotsmeta, STspots, STpropont\n",
    "\n",
    "def read_ontology():\n",
    "    ontology = pd.read_csv(ONTOLOGY_PATH)\n",
    "    ontology = ontology.drop([ontology.columns[5], ontology.columns[6]], axis=1)\n",
    "    ontology = ontology.fillna(-1)  #make root's parent -1\n",
    "\n",
    "    return ontology\n",
    "\n",
    "\n",
    "def getleaves(propontvox, ontology):\n",
    "    \"\"\"helper function to get only leaf brain areas\"\"\"\n",
    "    #leaves are brain areas in the ontology that never show up in the parent column\n",
    "    allareas = list(propontvox)\n",
    "    parents = list(ontology.parent)\n",
    "    for i in range(len(parents)): #convert parents from float to int, ids are ints\n",
    "        parents[i] = int(parents[i])\n",
    "    \n",
    "    #remove parents from all areas\n",
    "    leaves = []\n",
    "    for area in allareas:\n",
    "        if int(area) not in parents:\n",
    "            leaves.append(area)\n",
    "    \n",
    "    print(\"number of leaf areas: %d\" %len(leaves))\n",
    "    return leaves\n",
    "\n",
    "def findoverlapareas(STonto, propontvox, ontology):\n",
    "    \"\"\"find leaf brain areas overlapping between the two datasets\"\"\"\n",
    "    leafST = getleaves(STonto, ontology)\n",
    "    leafABA = getleaves(propontvox, ontology)\n",
    "\n",
    "    leafboth = [] \n",
    "    for i in range(len(leafABA)):\n",
    "        if leafABA[i] in leafST:\n",
    "            leafboth.append(leafABA[i])\n",
    "    \n",
    "    STonto = STonto.loc[:,leafboth]\n",
    "    propontvox = propontvox.loc[:,leafboth]\n",
    "    \n",
    "    return STonto, propontvox    \n",
    "\n",
    "def zscore(voxbrain):\n",
    "    \"\"\"zscore voxbrain or subsets of voxbrain (rows: voxels, cols: genes)\"\"\"\n",
    "    #z-score on whole data set before splitting into test and train\n",
    "    scaler = StandardScaler(copy=True, with_mean=True, with_std=True)\n",
    "    scaler.fit(voxbrain)\n",
    "    z_voxbrain = scaler.transform(voxbrain)\n",
    "    \n",
    "    #store z-scored voxbrain as pandas dataframe\n",
    "    z_voxbrain = pd.DataFrame(z_voxbrain)\n",
    "    z_voxbrain.columns = voxbrain.columns\n",
    "    z_voxbrain.index = voxbrain.index\n",
    "    \n",
    "    return z_voxbrain\n",
    "\n",
    "def analytical_auroc(featurevector, binarylabels):\n",
    "    \"\"\"analytical calculation of auroc\n",
    "       inputs: feature (mean rank of expression level), binary label (ctxnotctx)\n",
    "       returns: auroc\n",
    "    \"\"\"\n",
    "    #sort ctxnotctx binary labels by mean rank, aescending\n",
    "    s = sorted(zip(featurevector, binarylabels))\n",
    "    feature_sort, binarylabels_sort = map(list, zip(*s))\n",
    "\n",
    "    #get the sum of the ranks in feature vector corresponding to 1's in binary vector\n",
    "    sumranks = 0\n",
    "    for i in range(len(binarylabels_sort)):\n",
    "        if binarylabels_sort[i] == 1:\n",
    "            sumranks = sumranks + feature_sort[i]\n",
    "    \n",
    "    poslabels = binarylabels.sum()\n",
    "    neglabels = (len(binarylabels) - poslabels)\n",
    "    \n",
    "    auroc = ((sumranks/(neglabels*poslabels)) - ((poslabels+1)/(2*neglabels)))\n",
    "    \n",
    "    return auroc\n",
    "\n",
    "def getoverlapgenes(STspots, ABAvox):\n",
    "    ABAgenes = list(ABAvox)\n",
    "    STgenes = list(STspots)\n",
    "    \n",
    "    #get overlapping genes\n",
    "    overlap = []\n",
    "    for i in range(len(ABAgenes)):\n",
    "        if ABAgenes[i] in STgenes:\n",
    "            overlap.append(ABAgenes[i])\n",
    "    \n",
    "    print(\"number of overlapping genes: %d\" %len(overlap))\n",
    "    \n",
    "    #index datasets to keep only genes that are overlapping\n",
    "    STspots = STspots.loc[:,overlap]\n",
    "    ABAvox = ABAvox.loc[:,overlap]\n",
    "    \n",
    "    return STspots, ABAvox\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def filterproponto(sampleonto):\n",
    "    \"\"\"pre-processing for propogated ontology\"\"\"\n",
    "    #remove brain areas that don't have any samples\n",
    "    sampleonto_sums = sampleonto.apply(lambda col: col.sum(), axis=0)\n",
    "    sampleonto = sampleonto.loc[:,sampleonto_sums > 100] #100 to subset for dedicated multiclass\n",
    "    \n",
    "    return sampleonto"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Main"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-6-67adb9581260>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0;31m#filter brain areas for those that have at least x samples\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0mSTpropont\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfilterproponto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mSTpropont\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 11\u001b[0;31m \u001b[0mABApropont\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfilterproponto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mABApropont\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     12\u001b[0m \u001b[0;31m#filter brain areas for overlapping leaf areas\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0mSTpropont\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mABApropont\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfindoverlapareas\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mSTpropont\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mABApropont\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0montology\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-5-2298e2d81953>\u001b[0m in \u001b[0;36mfilterproponto\u001b[0;34m(sampleonto)\u001b[0m\n\u001b[1;32m     27\u001b[0m     \u001b[0;34m\"\"\"pre-processing for propogated ontology\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     28\u001b[0m     \u001b[0;31m#remove brain areas that don't have any samples\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 29\u001b[0;31m     \u001b[0msampleonto_sums\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msampleonto\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mcol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mcol\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     30\u001b[0m     \u001b[0msampleonto\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msampleonto\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msampleonto_sums\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m5\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;31m#greater than 5 becuase less is not enough for train/test split to have non-zero areas\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     31\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/frame.py\u001b[0m in \u001b[0;36mapply\u001b[0;34m(self, func, axis, broadcast, raw, reduce, result_type, args, **kwds)\u001b[0m\n\u001b[1;32m   6904\u001b[0m             \u001b[0mkwds\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6905\u001b[0m         )\n\u001b[0;32m-> 6906\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mop\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_result\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   6907\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   6908\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mapplymap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/apply.py\u001b[0m in \u001b[0;36mget_result\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    184\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapply_raw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    185\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 186\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapply_standard\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    187\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    188\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mapply_empty_result\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/apply.py\u001b[0m in \u001b[0;36mapply_standard\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    283\u001b[0m             \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    284\u001b[0m                 result = reduction.reduce(\n\u001b[0;32m--> 285\u001b[0;31m                     \u001b[0mvalues\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mf\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0maxis\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdummy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdummy\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlabels\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    286\u001b[0m                 )\n\u001b[1;32m    287\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_constructor_sliced\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mresult\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlabels\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/reduction.pyx\u001b[0m in \u001b[0;36mpandas._libs.reduction.reduce\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mpandas/_libs/reduction.pyx\u001b[0m in \u001b[0;36mpandas._libs.reduction.Reducer.get_result\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32m<ipython-input-5-2298e2d81953>\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(col)\u001b[0m\n\u001b[1;32m     27\u001b[0m     \u001b[0;34m\"\"\"pre-processing for propogated ontology\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     28\u001b[0m     \u001b[0;31m#remove brain areas that don't have any samples\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 29\u001b[0;31m     \u001b[0msampleonto_sums\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msampleonto\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapply\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mcol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mcol\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maxis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     30\u001b[0m     \u001b[0msampleonto\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msampleonto\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mloc\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0msampleonto_sums\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m5\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;31m#greater than 5 becuase less is not enough for train/test split to have non-zero areas\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     31\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/generic.py\u001b[0m in \u001b[0;36mstat_func\u001b[0;34m(self, axis, skipna, level, numeric_only, min_count, **kwargs)\u001b[0m\n\u001b[1;32m  11585\u001b[0m             \u001b[0mskipna\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mskipna\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m  11586\u001b[0m             \u001b[0mnumeric_only\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mnumeric_only\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m> 11587\u001b[0;31m             \u001b[0mmin_count\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmin_count\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m  11588\u001b[0m         )\n\u001b[1;32m  11589\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/series.py\u001b[0m in \u001b[0;36m_reduce\u001b[0;34m(self, op, name, axis, skipna, numeric_only, filter_type, **kwds)\u001b[0m\n\u001b[1;32m   4081\u001b[0m                 )\n\u001b[1;32m   4082\u001b[0m             \u001b[0;32mwith\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0merrstate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mall\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"ignore\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 4083\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdelegate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mskipna\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mskipna\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   4084\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   4085\u001b[0m         \u001b[0;31m# TODO(EA) dispatch to Index\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/nanops.py\u001b[0m in \u001b[0;36m_f\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     68\u001b[0m             \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     69\u001b[0m                 \u001b[0;32mwith\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0merrstate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minvalid\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"ignore\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 70\u001b[0;31m                     \u001b[0;32mreturn\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     71\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mValueError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     72\u001b[0m                 \u001b[0;31m# we want to transform an object array\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/nanops.py\u001b[0m in \u001b[0;36mnansum\u001b[0;34m(values, axis, skipna, min_count, mask)\u001b[0m\n\u001b[1;32m    493\u001b[0m     \"\"\"\n\u001b[1;32m    494\u001b[0m     values, mask, dtype, dtype_max, _ = _get_values(\n\u001b[0;32m--> 495\u001b[0;31m         \u001b[0mvalues\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mskipna\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfill_value\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmask\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmask\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    496\u001b[0m     )\n\u001b[1;32m    497\u001b[0m     \u001b[0mdtype_sum\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdtype_max\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/nanops.py\u001b[0m in \u001b[0;36m_get_values\u001b[0;34m(values, skipna, fill_value, fill_value_typ, mask)\u001b[0m\n\u001b[1;32m    279\u001b[0m         \u001b[0mfill\u001b[0m \u001b[0mvalue\u001b[0m \u001b[0mused\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    280\u001b[0m     \"\"\"\n\u001b[0;32m--> 281\u001b[0;31m     \u001b[0mmask\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_maybe_get_mask\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mskipna\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmask\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    282\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    283\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mis_datetime64tz_dtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/nanops.py\u001b[0m in \u001b[0;36m_maybe_get_mask\u001b[0;34m(values, skipna, mask)\u001b[0m\n\u001b[1;32m    232\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    233\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mskipna\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 234\u001b[0;31m             \u001b[0mmask\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0misna\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    235\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    236\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mmask\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/dtypes/missing.py\u001b[0m in \u001b[0;36misna\u001b[0;34m(obj)\u001b[0m\n\u001b[1;32m    120\u001b[0m     \u001b[0mName\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mbool\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    121\u001b[0m     \"\"\"\n\u001b[0;32m--> 122\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0m_isna\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    123\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/dtypes/missing.py\u001b[0m in \u001b[0;36m_isna_new\u001b[0;34m(obj)\u001b[0m\n\u001b[1;32m    143\u001b[0m         ),\n\u001b[1;32m    144\u001b[0m     ):\n\u001b[0;32m--> 145\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0m_isna_ndarraylike\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    146\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mABCGeneric\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    147\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mobj\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_constructor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_data\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0misna\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfunc\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0misna\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/batch/lib/python3.7/site-packages/pandas/core/dtypes/missing.py\u001b[0m in \u001b[0;36m_isna_ndarraylike\u001b[0;34m(obj)\u001b[0m\n\u001b[1;32m    250\u001b[0m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mvalues\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mview\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"i8\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0miNaT\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    251\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 252\u001b[0;31m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0misnan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    253\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    254\u001b[0m     \u001b[0;31m# box\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "#CROSS\n",
    "\n",
    "#read in data\n",
    "ontology = read_ontology()\n",
    "ABAmeta, ABAvox, ABApropont = read_ABAdata()\n",
    "\n",
    "STmeta, STspots, STpropont = read_STdata()\n",
    "\n",
    "#filter brain areas for those that have at least x samples\n",
    "STpropont = filterproponto(STpropont)\n",
    "ABApropont = filterproponto(ABApropont)\n",
    "#filter brain areas for overlapping leaf areas\n",
    "STpropont, ABApropont = findoverlapareas(STpropont, ABApropont, ontology)\n",
    "\n",
    "#keep only genes that are overlapping between the two datasets\n",
    "STspots, ABAvox = getoverlapgenes(STspots, ABAvox)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of leaf areas: 65\n"
     ]
    }
   ],
   "source": [
    "#JUST ST\n",
    "ontology = read_ontology()\n",
    "#pre-processing\n",
    "STspotsmeta, STspots, STpropont = read_STdata()\n",
    "STpropont = filterproponto(STpropont)\n",
    "STspots = STspots.astype('float64') #convert int to float for z-scoring\n",
    "#get leaf areas\n",
    "leaves = getleaves(STpropont, ontology)\n",
    "STpropont = STpropont.loc[STspotsmeta.id.isin(leaves),leaves] #subset prop onto for leaf areas\n",
    "STspots = STspots.loc[STspotsmeta.id.isin(leaves),:] #subset data for samples from leaves\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "number of leaf areas: 139\n"
     ]
    }
   ],
   "source": [
    "#JUST ABA\n",
    "#read in data\n",
    "ontology = read_ontology()\n",
    "ABAmeta, ABAvox, ABApropont = read_ABAdata()\n",
    "#pre-processing\n",
    "ABApropont = filterproponto(ABApropont)\n",
    "#get leaf ares only\n",
    "leaves = getleaves(ABApropont,ontology)\n",
    "ABApropont = ABApropont.loc[ABAmeta.ids.isin(leaves),leaves]\n",
    "ABAvox = ABAvox.loc[ABAmeta.ids.isin(leaves),:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# OVR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def applyLASSO(Xtrain, Xtest, ytrain, ytest):\n",
    "    \"\"\"apply LASSO regression\"\"\"\n",
    "    #lasso_reg = Lasso(alpha=0.01, max_iter=10000)\n",
    "    lasso_reg = LinearRegression()\n",
    "    lasso_reg.fit(Xtrain, ytrain)\n",
    "    \n",
    "    #train\n",
    "    predictions_train = lasso_reg.predict(Xtrain)\n",
    "    #auroc_train = analytical_auroc(sp.stats.mstats.rankdata(predictions_train), ytrain)\n",
    "    #test\n",
    "    predictions_test = lasso_reg.predict(Xtest)\n",
    "    #auroc_test = analytical_auroc(sp.stats.mstats.rankdata(predictions_test), ytest)\n",
    "    \n",
    "    return sp.stats.mstats.rankdata(predictions_train), sp.stats.mstats.rankdata(predictions_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "col 0\n",
      "col 1\n",
      "col 2\n",
      "col 3\n",
      "col 4\n",
      "col 5\n"
     ]
    }
   ],
   "source": [
    "mod_propont = leafSTpropont\n",
    "mod_data = leafSTspots\n",
    "#cross_propont = ABApropont\n",
    "#cross_data = ABAvox\n",
    "\n",
    "trainpreds = pd.DataFrame(columns=list(mod_propont))\n",
    "testpreds = pd.DataFrame(columns=list(mod_propont))\n",
    "    \n",
    "areas = list(mod_propont)\n",
    "#for each column, brain area\n",
    "for i in range(mod_propont.shape[1]):\n",
    "    print(\"col %d\" %i)\n",
    "    area1 = areas[i]\n",
    "\n",
    "    #get binary label vectors\n",
    "    ylabels = mod_propont.loc[:, area1]\n",
    "\n",
    "    #split train test for X data and y labels\n",
    "    #split data function is seeded so all will split the same way\n",
    "    Xtrain, Xtest, ytrain, ytest = train_test_split(mod_data, ylabels, test_size=0.5,\\\n",
    "                                                    random_state=42, shuffle=True,\\\n",
    "                                                    stratify=ylabels)\n",
    "    #z-score train and test folds\n",
    "    zXtrain = zscore(Xtrain)\n",
    "    zXtest = zscore(Xtest)\n",
    "\n",
    "    #fit LASSO on train set for 1 v all\n",
    "    currpred_train,currpred_test = applyLASSO(zXtrain, zXtest, ytrain, ytest)\n",
    "    #key = \"%s,%s\" %(str(i))\n",
    "    trainpreds[area1] = currpred_train\n",
    "    testpreds[area1] = currpred_test\n",
    "    #break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainpreds.to_csv(\"trainpreds_040321.csv\", sep=',',header=True,index=False)\n",
    "testpreds.to_csv(\"testpreds_040321.csv\", sep=',',header=True,index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainpreds = pd.read_csv(\"trainpreds_040321.csv\", sep=',')\n",
    "testpreds = pd.read_csv(\"testpreds_040321.csv\", sep=',')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "#rank rows\n",
    "trainranks = sp.stats.mstats.rankdata(trainpreds.to_numpy(),axis=1)\n",
    "testranks = sp.stats.mstats.rankdata(testpreds.to_numpy(),axis=1)\n",
    "\n",
    "#convert back to pandas\n",
    "trainranks = pd.DataFrame(trainranks, index=trainpreds.index, columns=trainpreds.columns)\n",
    "testranks = pd.DataFrame(testranks, index=testpreds.index, columns=testpreds.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "col 0\n",
      "col 5\n",
      "col 10\n",
      "col 15\n",
      "col 20\n",
      "col 25\n",
      "col 30\n",
      "col 35\n",
      "col 40\n",
      "col 45\n",
      "col 50\n",
      "col 55\n",
      "col 60\n",
      "col 65\n",
      "col 70\n",
      "col 75\n",
      "col 80\n",
      "col 85\n",
      "col 90\n",
      "col 95\n",
      "col 100\n",
      "col 105\n",
      "col 110\n",
      "col 115\n",
      "col 120\n",
      "col 125\n",
      "col 130\n",
      "col 135\n",
      "col 140\n",
      "col 145\n",
      "col 150\n",
      "col 155\n",
      "col 160\n",
      "col 165\n",
      "col 170\n",
      "col 175\n",
      "col 180\n",
      "col 185\n",
      "col 190\n",
      "col 195\n",
      "col 200\n",
      "col 205\n",
      "col 210\n",
      "col 215\n",
      "col 220\n",
      "col 225\n",
      "col 230\n",
      "col 235\n",
      "col 240\n",
      "col 245\n",
      "col 250\n",
      "col 255\n",
      "col 260\n",
      "col 265\n",
      "col 270\n",
      "col 275\n",
      "col 280\n",
      "col 285\n",
      "col 290\n",
      "col 295\n",
      "col 300\n",
      "col 305\n",
      "col 310\n",
      "col 315\n",
      "col 320\n",
      "col 325\n",
      "col 330\n",
      "col 335\n",
      "col 340\n",
      "col 345\n",
      "col 350\n",
      "col 355\n",
      "col 360\n",
      "col 365\n",
      "col 370\n",
      "col 375\n"
     ]
    }
   ],
   "source": [
    "#get aurocs\n",
    "trainauroc = {}\n",
    "testauroc = {}\n",
    "\n",
    "areas = list(mod_propont)\n",
    "#for each column, brain area\n",
    "for i in range(mod_propont.shape[1]):\n",
    "    if i %5 == 0:\n",
    "        print(\"col %d\" %i)\n",
    "    area1 = areas[i]\n",
    "\n",
    "    #get binary label vectors\n",
    "    ylabels = mod_propont.loc[:, area1]\n",
    "\n",
    "    #split train test for X data and y labels\n",
    "    #split data function is seeded so all will split the same way\n",
    "    Xtrain, Xtest, ytrain, ytest = train_test_split(mod_data, ylabels, test_size=0.5,\\\n",
    "                                                    random_state=42, shuffle=True,\\\n",
    "                                                    stratify=ylabels)\n",
    "\n",
    "    #get auroc of new prediction vector\n",
    "    trainauroc[area1] = analytical_auroc(sp.stats.mstats.rankdata(trainranks[area1]), ytrain)\n",
    "    testauroc[area1] = analytical_auroc(sp.stats.mstats.rankdata(testranks[area1]), ytest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'trainauroc' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-84-13c064b3316e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrainauroc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mxlabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"ST train AUROC\\nOVR Linear Regression\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mylabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"number of brain areas\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m0.98\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m120\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"mean = %.3f\"\u001b[0m \u001b[0;34m%\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmean\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrainauroc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;31m#plt.savefig(\"040521_1.pdf\", bbox_inches='tight')\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'trainauroc' is not defined"
     ]
    }
   ],
   "source": [
    "plt.hist(trainauroc.values())\n",
    "plt.xlabel(\"ST train AUROC\\nOVR Linear Regression\")\n",
    "plt.ylabel(\"number of brain areas\")\n",
    "plt.text(0.98,120,\"mean = %.3f\" %np.mean(list(trainauroc.values())))\n",
    "plt.savefig(\"040521_1.pdf\", bbox_inches='tight')\n",
    "plt.show()\n",
    "\n",
    "print(np.mean(list(trainauroc.values())))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6154376588641873\n"
     ]
    }
   ],
   "source": [
    "plt.hist(testauroc.values())\n",
    "plt.xlabel(\"ST test AUROC\\nOVR Linear Regression\")\n",
    "plt.ylabel(\"number of brain areas\")\n",
    "plt.text(0.85,100,\"mean = %.3f\" %np.mean(list(testauroc.values())))\n",
    "plt.savefig(\"040521_2.pdf\", bbox_inches='tight')\n",
    "plt.show()\n",
    "\n",
    "print(sp.mean(list(testauroc.values())))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dedicated multiclass classifier: KNN\n",
    "# ST"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "only for 65 brain areas>100 for speed of running"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#collapse STpropont so I can use it to stratify train/test split\n",
    "columns = STpropont.columns\n",
    "ycollapse = pd.Series(index = STpropont.index)\n",
    "for i in range(len(columns)):\n",
    "    currrows = np.where(STpropont[columns[i]] == 1.0)[0]\n",
    "    ycollapse.iloc[currrows] = columns[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#to get single y vector split\n",
    "Xtrain, Xtest, ytrain, ytest = train_test_split(STspots, ycollapse, test_size=0.5,\\\n",
    "                                                    random_state=42, shuffle=True,\\\n",
    "                                                    stratify=ycollapse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "#to get propont split on the same indicies\n",
    "null1, null2, ytrain2, ytest2 = train_test_split(STspots, STpropont, test_size=0.5,\\\n",
    "                                                    random_state=42, shuffle=True,\\\n",
    "                                                    stratify=ycollapse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#z-score\n",
    "zXtrain = zscore(Xtrain)\n",
    "zXtest = zscore(Xtest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "                     metric_params=None, n_jobs=-1, n_neighbors=5, p=2,\n",
       "                     weights='uniform')"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = KNeighborsClassifier(n_neighbors=5, weights='uniform', algorithm='auto', n_jobs=-1)\n",
    "model.fit(zXtrain, ytrain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "accuracy = model.score(zXtest.iloc[0:10,:], ytest.iloc[0:10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.1"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "testpreds = model.predict(zXtest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "#break predictions vector out\n",
    "testpredmatrix = pd.DataFrame(0, index=(range(len(testpreds))), columns = STpropont.columns)\n",
    "for i in range(len(testpredmatrix.columns)):\n",
    "    area = testpredmatrix.columns[i]\n",
    "    for j in range(len(testpreds)):\n",
    "        if testpreds[j] == area:\n",
    "            testpredmatrix.iloc[j,i] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "testaurocs = []\n",
    "for col in testpredmatrix.columns:\n",
    "    testaurocs.append(analytical_auroc(sp.stats.mstats.rankdata(testpredmatrix.loc[:,col]),ytest2.loc[:,col]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(testaurocs)\n",
    "plt.xlabel(\"test set AUROC\\nST\")\n",
    "plt.ylabel(\"brain areas\")\n",
    "plt.text(0.65, 43, \"mean = %.3f\" %np.mean(testaurocs),size=20)\n",
    "plt.savefig(\"051021_1.pdf\", bbox_inches='tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainpreds = model.predict(zXtrain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [],
   "source": [
    "#break predictions vector out\n",
    "trainpredmatrix = pd.DataFrame(0, index=(range(len(trainpreds))), columns = STpropont.columns)\n",
    "for i in range(len(trainpredmatrix.columns)):\n",
    "    area = trainpredmatrix.columns[i]\n",
    "    for j in range(len(trainpreds)):\n",
    "        if trainpreds[j] == area:\n",
    "            trainpredmatrix.iloc[j,i] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainaurocs = []\n",
    "for col in trainpredmatrix.columns:\n",
    "    trainaurocs.append(analytical_auroc(sp.stats.mstats.rankdata(trainpredmatrix.loc[:,col]),ytrain2.loc[:,col]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(trainaurocs)\n",
    "plt.xlabel(\"train set AUROC\\nST\")\n",
    "plt.ylabel(\"brain areas\")\n",
    "plt.text(0.75, 30, \"mean = %.3f\" %np.mean(trainaurocs),size=20)\n",
    "plt.savefig(\"051021_2.pdf\", bbox_inches='tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ABA"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "only for 139 brain areas>100 for speed of running"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "#collapse ABApropont so I can use it to stratify train/test split\n",
    "columns = ABApropont.columns\n",
    "ycollapse = pd.Series(index = ABApropont.index)\n",
    "for i in range(len(columns)):\n",
    "    currrows = np.where(ABApropont[columns[i]] == 1.0)[0]\n",
    "    ycollapse.iloc[currrows] = columns[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [],
   "source": [
    "#to get single y vector split\n",
    "Xtrain, Xtest, ytrain, ytest = train_test_split(ABAvox, ycollapse, test_size=0.5,\\\n",
    "                                                    random_state=42, shuffle=True,\\\n",
    "                                                    stratify=ycollapse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "#to get propont split on the same indicies\n",
    "null1, null2, ytrain2, ytest2 = train_test_split(ABAvox, ABApropont, test_size=0.5,\\\n",
    "                                                    random_state=42, shuffle=True,\\\n",
    "                                                    stratify=ycollapse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [],
   "source": [
    "#z-score\n",
    "zXtrain = zscore(Xtrain)\n",
    "zXtest = zscore(Xtest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "                     metric_params=None, n_jobs=-1, n_neighbors=5, p=2,\n",
       "                     weights='uniform')"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = KNeighborsClassifier(n_neighbors=5, weights='uniform', algorithm='auto', n_jobs=-1)\n",
    "model.fit(zXtrain, ytrain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "testpreds = model.predict(zXtest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [],
   "source": [
    "#break predictions vector out\n",
    "ABAtestpredmatrix = pd.DataFrame(0, index=(range(len(testpreds))), columns = ABApropont.columns)\n",
    "for i in range(len(ABAtestpredmatrix.columns)):\n",
    "    area = ABAtestpredmatrix.columns[i]\n",
    "    for j in range(len(testpreds)):\n",
    "        if testpreds[j] == area:\n",
    "            ABAtestpredmatrix.iloc[j,i] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [],
   "source": [
    "ABAtestaurocs = []\n",
    "for col in ABAtestpredmatrix.columns:\n",
    "    ABAtestaurocs.append(analytical_auroc(sp.stats.mstats.rankdata(ABAtestpredmatrix.loc[:,col]),ytest2.loc[:,col]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(ABAtestaurocs)\n",
    "plt.xlabel(\"test set AUROC\\nABA\")\n",
    "plt.ylabel(\"brain areas\")\n",
    "plt.text(0.9, 25, \"mean = %.3f\" %np.mean(ABAtestaurocs),size=20)\n",
    "plt.savefig(\"051021_3.pdf\", bbox_inches='tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ABAtrainpreds = model.predict(zXtrain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [],
   "source": [
    "#break predictions vector out\n",
    "ABAtrainpredmatrix = pd.DataFrame(0, index=(range(len(ABAtrainpreds))), columns = ABApropont.columns)\n",
    "for i in range(len(ABAtrainpredmatrix.columns)):\n",
    "    area = ABAtrainpredmatrix.columns[i]\n",
    "    for j in range(len(ABAtrainpreds)):\n",
    "        if ABAtrainpreds[j] == area:\n",
    "            ABAtrainpredmatrix.iloc[j,i] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [],
   "source": [
    "ABAtrainaurocs = []\n",
    "for col in ABAtrainpredmatrix.columns:\n",
    "    ABAtrainaurocs.append(analytical_auroc(sp.stats.mstats.rankdata(ABAtrainpredmatrix.loc[:,col]),ytrain2.loc[:,col]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(ABAtrainaurocs)\n",
    "plt.xlabel(\"train set AUROC\\nABA\")\n",
    "plt.ylabel(\"brain areas\")\n",
    "plt.text(0.95, 32, \"mean = %.3f\" %np.mean(ABAtrainaurocs),size=20)\n",
    "plt.savefig(\"051021_4.pdf\", bbox_inches='tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}