tests/noepistasisgaussianlikelihood_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# `NoEpistasisGaussianLikelihood` model\n",
    "Simulates some variants and analyzes them under the `globalepistasis.NoEpistasisGaussianLikelihood` model.\n",
    "\n",
    "This notebook is designed as a test, not for user documentation--and it uses some private methods that should not be used in real analyses."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Set up\n",
    "Import Python modules:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import copy\n",
    "import math\n",
    "import random\n",
    "import time\n",
    "import warnings\n",
    "\n",
    "import binarymap\n",
    "\n",
    "import numpy\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "from plotnine import *\n",
    "\n",
    "import scipy\n",
    "\n",
    "import dms_variants.globalepistasis\n",
    "import dms_variants.codonvarianttable\n",
    "import dms_variants.plotnine_themes\n",
    "import dms_variants.simulate\n",
    "from dms_variants.constants import CBPALETTE, CODONS_NOSTOP"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Parameters for simulations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "seed = 1  # random number seed\n",
    "genelength = 25  # gene length in codons\n",
    "bclen = 16  # length of barcodes\n",
    "libs = [\"lib_1\", \"lib_2\"]  # distinct libraries of gene\n",
    "variants_per_lib = 500 * genelength  # variants per library\n",
    "avgmuts = 2.5  # average codon mutations per variant\n",
    "variant_error_rate = 0  # rate variants are mis-sequenced\n",
    "depth = 1000 * variants_per_lib  # per-sample sequencing depth\n",
    "lib_uniformity = 5  # uniformity of library pre-selection\n",
    "noise = 0.0  # non-statistical noise in selections\n",
    "bottlenecks = {  # bottlenecks from pre- to post-selection\n",
    "    \"loose_bottle\": variants_per_lib * 1000,\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Seed random number generators:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "random.seed(seed)\n",
    "numpy.random.seed(seed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Suppress warnings that clutter output:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "warnings.simplefilter(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set plotnine theme:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "theme_set(dms_variants.plotnine_themes.theme_graygrid())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simulate variants\n",
    "Simulate wildtype gene sequence:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wildtype gene of 25 codons:\n",
      "AGATCCGTGATTCTGCGTGCTTACACCAACTCACGGGTGAAACGTGTAATCTTATGCAACAACGACTTACCTATC\n"
     ]
    }
   ],
   "source": [
    "geneseq = \"\".join(random.choices(CODONS_NOSTOP, k=genelength))\n",
    "print(f\"Wildtype gene of {genelength} codons:\\n{geneseq}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Simulate a `CodonVariantTable`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "variants = dms_variants.simulate.simulate_CodonVariantTable(\n",
    "    geneseq=geneseq,\n",
    "    bclen=bclen,\n",
    "    library_specs={\n",
    "        lib: {\"avgmuts\": avgmuts, \"nvariants\": variants_per_lib} for lib in libs\n",
    "    },\n",
    "    seed=seed,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Number of amino-acid mutations per variant in library:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 550x210 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "\n",
    "p = variants.plotNumMutsHistogram(\"aa\", samples=None, max_muts=8)\n",
    "p = p + theme(panel_grid_major_x=element_blank())  # no vertical grid lines\n",
    "_ = p.draw()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simulate variant counts\n",
    "We will simulate using a `SigmoidPhenotypeSimulator`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "phenosimulator = dms_variants.simulate.SigmoidPhenotypeSimulator(\n",
    "    geneseq,\n",
    "    seed=seed,\n",
    "    norm_weights=((0.4, -0.5, 0.7), (0.6, -5, 1.5)),\n",
    "    stop_effect=-9,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "But since we are using a non-epistatic (linear model), we will simulate variants according to the *latent phenotype* rather than their *observed phenotype* (since in a non-epistatic model these two are the same).\n",
    "First, look at the distribution of latent phenotypes for all single mutants (the dashed orange line is the latent phenotype of wildtype):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 350x250 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "\n",
    "p = phenosimulator.plotMutsHistogram(\"latentPhenotype\")\n",
    "_ = p.draw()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Our simulation function is then the exponential of the latent phenotype, as mutants as the functional scores represent the log enrichment:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def phenotype_func(x):\n",
    "    return math.exp(phenosimulator.latentPhenotype(x))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we simulate the counts, passing to the simulate function the latent rather than observed phenotype:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "counts = dms_variants.simulate.simulateSampleCounts(\n",
    "    variants=variants,\n",
    "    phenotype_func=phenotype_func,\n",
    "    variant_error_rate=variant_error_rate,\n",
    "    pre_sample={\"total_count\": depth, \"uniformity\": lib_uniformity},\n",
    "    pre_sample_name=\"pre-selection\",\n",
    "    post_samples={\n",
    "        name: {\"noise\": noise, \"total_count\": depth, \"bottleneck\": bottle}\n",
    "        for name, bottle in bottlenecks.items()\n",
    "    },\n",
    "    seed=seed,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Add the counts to the variant table:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "variants.add_sample_counts_df(counts)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Confirm expected selection on mutations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 520x360 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "\n",
    "p = variants.plotNumCodonMutsByType(\"all\")\n",
    "p = p + theme(panel_grid_major_x=element_blank())  # no vertical grid lines\n",
    "_ = p.draw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>library</th>\n",
       "      <th>sample</th>\n",
       "      <th>mutation_type</th>\n",
       "      <th>num_muts_count</th>\n",
       "      <th>count</th>\n",
       "      <th>number</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>lib_1</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>nonsynonymous</td>\n",
       "      <td>28133593</td>\n",
       "      <td>12500000</td>\n",
       "      <td>2.250687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>lib_1</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>synonymous</td>\n",
       "      <td>1516864</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.121349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>lib_1</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>stop</td>\n",
       "      <td>1465269</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.117222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>lib_1</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>nonsynonymous</td>\n",
       "      <td>8745167</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.699613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>lib_1</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>synonymous</td>\n",
       "      <td>1385003</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.110800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>lib_1</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>stop</td>\n",
       "      <td>170</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.000014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>lib_2</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>nonsynonymous</td>\n",
       "      <td>27899970</td>\n",
       "      <td>12500000</td>\n",
       "      <td>2.231998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>lib_2</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>synonymous</td>\n",
       "      <td>1507947</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.120636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>lib_2</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>stop</td>\n",
       "      <td>1474547</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.117964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>lib_2</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>nonsynonymous</td>\n",
       "      <td>8403904</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.672312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>lib_2</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>synonymous</td>\n",
       "      <td>1493006</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.119440</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>lib_2</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>stop</td>\n",
       "      <td>177</td>\n",
       "      <td>12500000</td>\n",
       "      <td>0.000014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>all libraries</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>nonsynonymous</td>\n",
       "      <td>56033563</td>\n",
       "      <td>25000000</td>\n",
       "      <td>2.241343</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>all libraries</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>synonymous</td>\n",
       "      <td>3024811</td>\n",
       "      <td>25000000</td>\n",
       "      <td>0.120992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>all libraries</td>\n",
       "      <td>pre-selection</td>\n",
       "      <td>stop</td>\n",
       "      <td>2939816</td>\n",
       "      <td>25000000</td>\n",
       "      <td>0.117593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>all libraries</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>nonsynonymous</td>\n",
       "      <td>17149071</td>\n",
       "      <td>25000000</td>\n",
       "      <td>0.685963</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>all libraries</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>synonymous</td>\n",
       "      <td>2878009</td>\n",
       "      <td>25000000</td>\n",
       "      <td>0.115120</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>all libraries</td>\n",
       "      <td>loose_bottle</td>\n",
       "      <td>stop</td>\n",
       "      <td>347</td>\n",
       "      <td>25000000</td>\n",
       "      <td>0.000014</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          library         sample  mutation_type  num_muts_count     count  \\\n",
       "0           lib_1  pre-selection  nonsynonymous        28133593  12500000   \n",
       "1           lib_1  pre-selection     synonymous         1516864  12500000   \n",
       "2           lib_1  pre-selection           stop         1465269  12500000   \n",
       "3           lib_1   loose_bottle  nonsynonymous         8745167  12500000   \n",
       "4           lib_1   loose_bottle     synonymous         1385003  12500000   \n",
       "5           lib_1   loose_bottle           stop             170  12500000   \n",
       "6           lib_2  pre-selection  nonsynonymous        27899970  12500000   \n",
       "7           lib_2  pre-selection     synonymous         1507947  12500000   \n",
       "8           lib_2  pre-selection           stop         1474547  12500000   \n",
       "9           lib_2   loose_bottle  nonsynonymous         8403904  12500000   \n",
       "10          lib_2   loose_bottle     synonymous         1493006  12500000   \n",
       "11          lib_2   loose_bottle           stop             177  12500000   \n",
       "12  all libraries  pre-selection  nonsynonymous        56033563  25000000   \n",
       "13  all libraries  pre-selection     synonymous         3024811  25000000   \n",
       "14  all libraries  pre-selection           stop         2939816  25000000   \n",
       "15  all libraries   loose_bottle  nonsynonymous        17149071  25000000   \n",
       "16  all libraries   loose_bottle     synonymous         2878009  25000000   \n",
       "17  all libraries   loose_bottle           stop             347  25000000   \n",
       "\n",
       "      number  \n",
       "0   2.250687  \n",
       "1   0.121349  \n",
       "2   0.117222  \n",
       "3   0.699613  \n",
       "4   0.110800  \n",
       "5   0.000014  \n",
       "6   2.231998  \n",
       "7   0.120636  \n",
       "8   0.117964  \n",
       "9   0.672312  \n",
       "10  0.119440  \n",
       "11  0.000014  \n",
       "12  2.241343  \n",
       "13  0.120992  \n",
       "14  0.117593  \n",
       "15  0.685963  \n",
       "16  0.115120  \n",
       "17  0.000014  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "variants.numCodonMutsByType(\"all\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 700x280 with 6 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "\n",
    "p = variants.plotMutFreqs(\"all\", \"aa\")\n",
    "p = p + theme(panel_grid_major_x=element_blank())  # no vertical grid lines\n",
    "_ = p.draw()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get functional scores\n",
    "Get the functional scores:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "func_scores = variants.func_scores(\"pre-selection\", libraries=variants.libraries)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Classify the functional scores:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "func_scores = variants.classifyVariants(func_scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use `plotnine` to plot distribution of functional scores:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 550x200 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "\n",
    "p = (\n",
    "    ggplot(func_scores, aes(\"variant_class\", \"func_score\"))\n",
    "    + geom_violin(aes(fill=\"variant_class\"))\n",
    "    + ylab(\"functional score\")\n",
    "    + xlab(\"\")\n",
    "    + facet_grid(\"post_sample ~ library\")\n",
    "    + theme(\n",
    "        figure_size=(2.75 * len(libs), 2 * len(bottlenecks)),\n",
    "        axis_text_x=element_text(angle=90),\n",
    "        panel_grid_major_x=element_blank(),  # no vertical grid lines\n",
    "    )\n",
    "    + scale_fill_manual(values=CBPALETTE[1:])\n",
    ")\n",
    "_ = p.draw()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Make a binary map for the variants:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length of binary map is 500\n"
     ]
    }
   ],
   "source": [
    "binarymap = binarymap.BinaryMap(func_scores)\n",
    "print(f\"Length of binary map is {binarymap.binarylength}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now initialize the `NoEpistasisGaussianLikelihood` model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = dms_variants.globalepistasis.NoEpistasisGaussianLikelihood(binarymap)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Look at some model attributes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "502"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.nparams"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.likelihood_calc_params_dict[\"epistasis_HOC\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-447718"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(model.loglik)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.latent_phenotype_wt()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check the analytical gradient versus numerical approximation, and make sure the difference is small compared to total gradient:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "err = scipy.optimize.check_grad(\n",
    "    model._loglik_by_allparams, model._dloglik_by_allparams, model._allparams\n",
    ")\n",
    "tot = math.sqrt(sum(model._dloglik_by_allparams(model._allparams) ** 2))\n",
    "if err / tot > 1e-5:\n",
    "    raise ValueError(\n",
    "        f\"Numerical and analytical gradients differ: {err}, {tot}, {err / tot}\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initial least squares fitting\n",
    "Perform initial least squares optimization (normally this is actually performed by the `fit` method):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "model._fit_latent_leastsquares()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Re-check the gradients:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "err = scipy.optimize.check_grad(\n",
    "    model._loglik_by_allparams, model._dloglik_by_allparams, model._allparams\n",
    ")\n",
    "tot = math.sqrt(sum(model._dloglik_by_allparams(model._allparams) ** 2))\n",
    "if err / tot > 1e-3:\n",
    "    raise ValueError(\n",
    "        f\"Numerical and analytical gradients differ: {err}, {tot}, {err / tot}\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-60249"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(model.loglik)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R = 0.837, P = 0\n"
     ]
    }
   ],
   "source": [
    "r, p = scipy.stats.pearsonr(\n",
    "    binarymap.func_scores,\n",
    "    model.phenotypes_frombinary(binarymap.binary_variants, \"latent\"),\n",
    ")\n",
    "print(f\"R = {r:.3f}, P = {p:.2g}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "\n",
    "df = pd.DataFrame(\n",
    "    {\n",
    "        \"actual\": binarymap.func_scores,\n",
    "        \"predicted\": model.phenotypes_frombinary(binarymap.binary_variants, \"latent\"),\n",
    "    }\n",
    ")\n",
    "\n",
    "_ = (ggplot(df, aes(\"actual\", \"predicted\")) + geom_bin2d()).draw()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epistasis_HOC: 5.816\n"
     ]
    }
   ],
   "source": [
    "for param, val in model.likelihood_calc_params_dict.items():\n",
    "    print(f\"{param}: {val:.3f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-3.203"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(model.latent_phenotype_wt(), 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fit the full model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now fit the model by maximum likelihood.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting took 0.105 seconds\n"
     ]
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "start = time.time()\n",
    "optres = model.fit()\n",
    "print(f\"Fitting took {time.time() - start:.3f} seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-59809"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(model.loglik)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "epistasis_HOC: 4.837\n"
     ]
    }
   ],
   "source": [
    "for param, val in model.likelihood_calc_params_dict.items():\n",
    "    print(f\"{param}: {val:.3f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-2.49"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(model.latent_phenotype_wt(), 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R = 0.837, P = 0\n"
     ]
    }
   ],
   "source": [
    "r, p = scipy.stats.pearsonr(\n",
    "    binarymap.func_scores,\n",
    "    model.phenotypes_frombinary(binarymap.binary_variants, \"latent\"),\n",
    ")\n",
    "print(f\"R = {r:.3f}, P = {p:.2g}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "\n",
    "df = pd.DataFrame(\n",
    "    {\n",
    "        \"actual\": binarymap.func_scores,\n",
    "        \"predicted\": model.phenotypes_frombinary(binarymap.binary_variants, \"latent\"),\n",
    "    }\n",
    ")\n",
    "\n",
    "_ = (ggplot(df, aes(\"actual\", \"predicted\")) + geom_bin2d()).draw()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fit model with **two** latent phenotypes\n",
    "Initialize model with 2 latent phenotypes, using the prior 1-latent phenotype model.\n",
    "Note that for a non-epistatic model, a multi-latent model should be identical to a single-latent model, so some of the tests below will confirm that:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2 = dms_variants.globalepistasis.NoEpistasisGaussianLikelihood(\n",
    "    binarymap=binarymap, n_latent_phenotypes=2, model_one_less_latent=model\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Number of parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1003"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.nparams"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some initial parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4.84"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(model2.likelihood_calc_params_dict[\"epistasis_HOC\"], 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-59809"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "round(model2.loglik)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 -2.49\n",
      "2 0.0\n"
     ]
    }
   ],
   "source": [
    "for k in range(1, model2.n_latent_phenotypes + 1):\n",
    "    print(k, round(model2.latent_phenotype_wt(k=k), 2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check the analytical gradient versus numerical approximation, and make sure the difference is small compared to total gradient.\n",
    "In the test below, we're actually pretty tolerant since the model is at the max so the total gradient will be small:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "err = scipy.optimize.check_grad(\n",
    "    model2._loglik_by_allparams, model2._dloglik_by_allparams, model2._allparams\n",
    ")\n",
    "tot = math.sqrt(sum(model2._dloglik_by_allparams(model2._allparams) ** 2))\n",
    "if err / tot > 1e-1:\n",
    "    raise ValueError(\n",
    "        f\"Numerical and analytical gradients differ: {err}, {tot}, {err / tot}\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now re-test the gradients with a more stringent tolerance after \"un-fitting\" the model away from the maximum:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "unfit_model2 = copy.deepcopy(model2)\n",
    "unfit_model2._latenteffects = numpy.zeros(unfit_model2._latenteffects.shape)\n",
    "\n",
    "err = scipy.optimize.check_grad(\n",
    "    unfit_model2._loglik_by_allparams,\n",
    "    unfit_model2._dloglik_by_allparams,\n",
    "    unfit_model2._allparams,\n",
    ")\n",
    "tot = math.sqrt(sum(unfit_model2._dloglik_by_allparams(unfit_model2._allparams) ** 2))\n",
    "if err / tot > 1e-5:\n",
    "    raise ValueError(\n",
    "        f\"Numerical and analytical gradients differ: {err}, {tot}, {err / tot}\"\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fit the model by maximum likelihood:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting took 0.016 seconds\n"
     ]
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "start = time.time()\n",
    "optres = model2.fit()\n",
    "print(f\"Fitting took {time.time() - start:.3f} seconds\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For a non-epistatic model, the multi-latent phenotype model should be the same as the single latent phenotype in terms of likelihood:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not numpy.allclose(model.loglik, model2.loglik, atol=0.1):\n",
    "    raise ValueError(f\"log likelihoods differ: {model.loglik} vs {model2.loglik}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The one-latent phenotype wildtype latent should be the same as the sum of the latent phenotypes of multi-latent model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not numpy.allclose(\n",
    "    model.latent_phenotype_wt(),\n",
    "    sum(\n",
    "        model2.latent_phenotype_wt(k) for k in range(1, model2.n_latent_phenotypes + 1)\n",
    "    ),\n",
    "    atol=0.1,\n",
    "):\n",
    "    raise ValueError(\"wt latent phenotypes differ\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are no epistasis function parameters for a no-epistasis model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{}"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.epistasis_func_params_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The likelihood calculation parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "assert numpy.allclose(\n",
    "    model2.likelihood_calc_params_dict[\"epistasis_HOC\"], 4.84, atol=0.05\n",
    "), model2.likelihood_calc_params_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The number of latent effects should be twice the number of mutations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "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>mutation</th>\n",
       "      <th>latent_effect</th>\n",
       "      <th>latent_phenotype_number</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>R1*</td>\n",
       "      <td>-4.167662</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>R1A</td>\n",
       "      <td>0.173918</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>R1C</td>\n",
       "      <td>-1.183262</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>R1D</td>\n",
       "      <td>-4.572701</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>R1E</td>\n",
       "      <td>-3.082266</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  mutation  latent_effect  latent_phenotype_number\n",
       "0      R1*      -4.167662                        1\n",
       "1      R1A       0.173918                        1\n",
       "2      R1C      -1.183262                        1\n",
       "3      R1D      -4.572701                        1\n",
       "4      R1E      -3.082266                        1"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "model2.latent_effects_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "if model2.latent_effects_df.shape != (genelength * 20 * model2.n_latent_phenotypes, 3):\n",
    "    raise ValueError(f\"unexpected shape for latent effects\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For a no-epistasis model, the sum at a site of latent effects for a multi-latent phenotype model should equal the latent effects for a single-latent phenotype model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    model2.latent_effects_df.groupby(\"mutation\")\n",
    "    .aggregate(multi_latent_sum=pd.NamedAgg(\"latent_effect\", \"sum\"))\n",
    "    .reset_index()\n",
    "    .merge(model.latent_effects_df, on=\"mutation\")\n",
    "    .assign(\n",
    "        nearly_equal=lambda x: abs(x[\"multi_latent_sum\"] - x[\"latent_effect\"]) < 5e-2\n",
    "    )[\"nearly_equal\"]\n",
    "    .all()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Look at the phenotypes; there should be one for each latent phenotype:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "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>aa_substitutions</th>\n",
       "      <th>func_score</th>\n",
       "      <th>func_score_var</th>\n",
       "      <th>latent_phenotype_1</th>\n",
       "      <th>latent_phenotype_2</th>\n",
       "      <th>observed_phenotype</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td></td>\n",
       "      <td>0.023237</td>\n",
       "      <td>0.000643</td>\n",
       "      <td>-2.490396</td>\n",
       "      <td>-0.000116</td>\n",
       "      <td>-2.490512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>S11C N20G P24A</td>\n",
       "      <td>-15.140211</td>\n",
       "      <td>4.163345</td>\n",
       "      <td>-10.170586</td>\n",
       "      <td>-0.000100</td>\n",
       "      <td>-10.170686</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>R1P L18N</td>\n",
       "      <td>-5.984570</td>\n",
       "      <td>0.007924</td>\n",
       "      <td>-6.582197</td>\n",
       "      <td>-0.000184</td>\n",
       "      <td>-6.582381</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>I4A</td>\n",
       "      <td>-5.900341</td>\n",
       "      <td>0.007905</td>\n",
       "      <td>-5.391154</td>\n",
       "      <td>-0.000187</td>\n",
       "      <td>-5.391341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>L5G Y8R R15N L23Q</td>\n",
       "      <td>-15.038653</td>\n",
       "      <td>4.163389</td>\n",
       "      <td>-15.777580</td>\n",
       "      <td>-0.000174</td>\n",
       "      <td>-15.777754</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    aa_substitutions  func_score  func_score_var  latent_phenotype_1  \\\n",
       "0                       0.023237        0.000643           -2.490396   \n",
       "1     S11C N20G P24A  -15.140211        4.163345          -10.170586   \n",
       "2           R1P L18N   -5.984570        0.007924           -6.582197   \n",
       "3                I4A   -5.900341        0.007905           -5.391154   \n",
       "4  L5G Y8R R15N L23Q  -15.038653        4.163389          -15.777580   \n",
       "\n",
       "   latent_phenotype_2  observed_phenotype  \n",
       "0           -0.000116           -2.490512  \n",
       "1           -0.000100          -10.170686  \n",
       "2           -0.000184           -6.582381  \n",
       "3           -0.000187           -5.391341  \n",
       "4           -0.000174          -15.777754  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "model2.phenotypes_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['aa_substitutions',\n",
       " 'func_score',\n",
       " 'func_score_var',\n",
       " 'latent_phenotype_1',\n",
       " 'latent_phenotype_2',\n",
       " 'observed_phenotype']"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.phenotypes_df.columns.tolist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The observed phenotypes should be essentially the same for the one- and two-latent phenotype models:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    model2.phenotypes_df.rename(\n",
    "        columns={\"observed_phenotype\": \"multi_latent_observed_phenotype\"}\n",
    "    )\n",
    "    .merge(model.phenotypes_df)\n",
    "    .assign(\n",
    "        nearly_equal=lambda x: abs(\n",
    "            x[\"multi_latent_observed_phenotype\"] - x[\"observed_phenotype\"]\n",
    "        )\n",
    "        < 5e-2\n",
    "    )[\"nearly_equal\"]\n",
    "    .all()\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  mutation wildtype  site mutant  effect\n",
      "0      V3*        V     3      *   -2.31\n",
      "1      S2Y        S     2      Y   -2.26\n",
      "2      R6W        R     6      W   -2.15\n",
      "3      L5*        L     5      *   -2.12\n",
      "4     N20*        N    20      *   -2.08\n",
      "  mutation wildtype  site mutant  effect\n",
      "0     S11M        S    11      M   -7.76\n",
      "1     R12M        R    12      M   -7.60\n",
      "2     C19W        C    19      W   -6.30\n",
      "3     D22W        D    22      W   -4.88\n",
      "4     L18G        L    18      G   -3.76\n"
     ]
    }
   ],
   "source": [
    "# NBVAL_IGNORE_OUTPUT\n",
    "for k in range(1, model2.n_latent_phenotypes + 1):\n",
    "    print(model2.single_mut_effects(phenotype=\"latent\", k=k).head().round(2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.0"
  },
  "toc": {
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "toc_cell": false,
   "toc_position": {
    "height": "1033.71px",
    "left": "0px",
    "right": "1423.47px",
    "top": "110.284px",
    "width": "211.989px"
   },
   "toc_section_display": "block",
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}