20.umap-transform.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Convert STRING to a 2d embedding using UMAP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import umap\n",
    "import scipy.sparse\n",
    "import pandas as pd\n",
    "\n",
    "from utils import get_string_dataset_path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>index</th>\n",
       "      <th>protein_external_id</th>\n",
       "      <th>preferred_name</th>\n",
       "      <th>protein_size</th>\n",
       "      <th>annotation</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>19564</th>\n",
       "      <td>19564</td>\n",
       "      <td>9606.ENSP00000485675</td>\n",
       "      <td>ENSG00000280116</td>\n",
       "      <td>84</td>\n",
       "      <td>annotation not available</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19565</th>\n",
       "      <td>19565</td>\n",
       "      <td>9606.ENSP00000485678</td>\n",
       "      <td>OR6Q1</td>\n",
       "      <td>317</td>\n",
       "      <td>Olfactory receptor 6Q1; Odorant receptor; Olfa...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       index   protein_external_id   preferred_name  protein_size  \\\n",
       "19564  19564  9606.ENSP00000485675  ENSG00000280116            84   \n",
       "19565  19565  9606.ENSP00000485678            OR6Q1           317   \n",
       "\n",
       "                                              annotation  \n",
       "19564                           annotation not available  \n",
       "19565  Olfactory receptor 6Q1; Odorant receptor; Olfa...  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Show the head of protein info\n",
    "gene_path = get_string_dataset_path(\"protein.info\")\n",
    "gene_df = pd.read_csv(gene_path, sep='\\t')\n",
    "gene_df = gene_df.reset_index()\n",
    "gene_df.tail(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 734 ms, sys: 35.2 ms, total: 770 ms\n",
      "Wall time: 801 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "matrix = scipy.sparse.load_npz(\"data/score-matrices/combined_score.sparse.npz\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "reducer = umap.UMAP(\n",
    "    n_components=2,\n",
    "    random_state=0,\n",
    "    init=\"random\", # https://github.com/lmcinnes/umap/issues/381\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 29min 44s, sys: 3.06 s, total: 29min 48s\n",
      "Wall time: 29min 49s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "mapper = reducer.fit(matrix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>index</th>\n",
       "      <th>protein_external_id</th>\n",
       "      <th>preferred_name</th>\n",
       "      <th>protein_size</th>\n",
       "      <th>annotation</th>\n",
       "      <th>umap_0</th>\n",
       "      <th>umap_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>9606.ENSP00000000233</td>\n",
       "      <td>ARF5</td>\n",
       "      <td>180</td>\n",
       "      <td>ADP-ribosylation factor 5; GTP-binding protein...</td>\n",
       "      <td>7.348941</td>\n",
       "      <td>-0.685039</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>9606.ENSP00000000412</td>\n",
       "      <td>M6PR</td>\n",
       "      <td>277</td>\n",
       "      <td>Cation-dependent mannose-6-phosphate receptor;...</td>\n",
       "      <td>-1.427780</td>\n",
       "      <td>3.857086</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   index   protein_external_id preferred_name  protein_size  \\\n",
       "0      0  9606.ENSP00000000233           ARF5           180   \n",
       "1      1  9606.ENSP00000000412           M6PR           277   \n",
       "\n",
       "                                          annotation    umap_0    umap_1  \n",
       "0  ADP-ribosylation factor 5; GTP-binding protein...  7.348941 -0.685039  \n",
       "1  Cation-dependent mannose-6-phosphate receptor;... -1.427780  3.857086  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create embedding dataframe\n",
    "embedding_df = pd.DataFrame(mapper.transform(matrix)).add_prefix(\"umap_\")\n",
    "embedding_df = pd.concat([gene_df, embedding_df], axis=\"columns\")\n",
    "embedding_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "embedding_df.to_csv(\"data/embeddings/umap-2d.csv.xz\", sep=\"\\t\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plot embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import umap.plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'.  Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f92255e7640>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "umap.plot.points(mapper)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"display: table;\"><div style=\"display: table-row;\"><div style=\"display: table-cell;\"><b title=\"bokeh.models.renderers.GlyphRenderer\">GlyphRenderer</b>(</div><div style=\"display: table-cell;\">id&nbsp;=&nbsp;'1036', <span id=\"1039\" style=\"cursor: pointer;\">&hellip;)</span></div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">data_source&nbsp;=&nbsp;ColumnDataSource(id='1001', ...),</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">glyph&nbsp;=&nbsp;Circle(id='1034', ...),</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">hover_glyph&nbsp;=&nbsp;None,</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">js_event_callbacks&nbsp;=&nbsp;{},</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">js_property_callbacks&nbsp;=&nbsp;{},</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">level&nbsp;=&nbsp;'glyph',</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">muted&nbsp;=&nbsp;False,</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">muted_glyph&nbsp;=&nbsp;None,</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">name&nbsp;=&nbsp;None,</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">nonselection_glyph&nbsp;=&nbsp;Circle(id='1035', ...),</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">selection_glyph&nbsp;=&nbsp;None,</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">subscribed_events&nbsp;=&nbsp;[],</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">tags&nbsp;=&nbsp;[],</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">view&nbsp;=&nbsp;CDSView(id='1037', ...),</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">visible&nbsp;=&nbsp;True,</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">x_range_name&nbsp;=&nbsp;'default',</div></div><div class=\"1038\" style=\"display: none;\"><div style=\"display: table-cell;\"></div><div style=\"display: table-cell;\">y_range_name&nbsp;=&nbsp;'default')</div></div></div>\n",
       "<script>\n",
       "(function() {\n",
       "  var expanded = false;\n",
       "  var ellipsis = document.getElementById(\"1039\");\n",
       "  ellipsis.addEventListener(\"click\", function() {\n",
       "    var rows = document.getElementsByClassName(\"1038\");\n",
       "    for (var i = 0; i < rows.length; i++) {\n",
       "      var el = rows[i];\n",
       "      el.style.display = expanded ? \"none\" : \"table-row\";\n",
       "    }\n",
       "    ellipsis.innerHTML = expanded ? \"&hellip;)\" : \"&lsaquo;&lsaquo;&lsaquo;\";\n",
       "    expanded = !expanded;\n",
       "  });\n",
       "})();\n",
       "</script>\n"
      ],
      "text/plain": [
       "GlyphRenderer(id='1036', ...)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# based on example at https://umap-learn.readthedocs.io/en/latest/basic_usage.html\n",
    "import bokeh.plotting\n",
    "import bokeh.models\n",
    "import bokeh.palettes\n",
    "\n",
    "datasource = bokeh.models.ColumnDataSource(embedding_df)\n",
    "\n",
    "plot_title = 'UMAP projection of STRING proteins'\n",
    "figure = bokeh.plotting.figure(\n",
    "    title=plot_title,\n",
    "    plot_width=600,\n",
    "    plot_height=600,\n",
    "    tools='box_zoom, pan, wheel_zoom, reset'\n",
    ")\n",
    "\n",
    "hover_tool = bokeh.models.HoverTool(\n",
    "    tooltips=[\n",
    "        ('symbol', '<i>@preferred_name</i>'),\n",
    "        ('protein_size',  '@protein_size{,}'),\n",
    "        ('info', '@annotation'),\n",
    "    ],\n",
    ")\n",
    "figure.add_tools(hover_tool)\n",
    "\n",
    "figure.circle(\n",
    "    x='umap_0',\n",
    "    y='umap_1',\n",
    "    source=datasource,\n",
    "    line_alpha=0.6,\n",
    "    fill_alpha=0.6,\n",
    "    radius=0.01\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Uncomment below to display plot in notebook\n",
    "# bokeh.plotting.output_notebook()\n",
    "# bokeh.plotting.show(plot_figure)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/user/jupyter/data/embeddings/umap-2d-plot.html'"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import bokeh.resources\n",
    "bokeh.plotting.save(\n",
    "    obj=figure,\n",
    "    resources=bokeh.resources.CDN,\n",
    "    title=plot_title,\n",
    "    filename=\"data/embeddings/umap-2d-plot.html\",\n",
    ")"
   ]
  }
 ],
 "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}