deepmind-research/enformer/enformer-example.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "enformer-example.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rb_ShvB9E8yM"
      },
      "source": [
        "Copyright 2021 DeepMind Technologies Limited\n",
        "\n",
        "Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "you may not use this file except in compliance with the License.\n",
        "You may obtain a copy of the License at\n",
        "\n",
        "     https://www.apache.org/licenses/LICENSE-2.0\n",
        "\n",
        "Unless required by applicable law or agreed to in writing, software\n",
        "distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "See the License for the specific language governing permissions and\n",
        "limitations under the License."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kXQjDxgdwUmW"
      },
      "source": [
        "This colab showcases the usage of the Enformer model published in\n",
        "\n",
        "**\"Effective gene expression prediction from sequence by integrating long-range interactions\"**\n",
        "\n",
        "Žiga Avsec1*, Vikram Agarwal2,4, Daniel Visentin1,4, Joseph R. Ledsam1,3, Agnieszka Grabska-Barwinska1, Kyle R. Taylor1, Yannis Assael1, John Jumper1, Pushmeet Kohli1*, David R. Kelley2*\n",
        "\n",
        "- 1 DeepMind, London, UK\n",
        "- 2 Calico Life Sciences, South San Francisco, CA, USA\n",
        "- 3 Google, Tokyo, Japan\n",
        "- 4 These authors contributed equally.\n",
        "- `*` correspondence: avsec@google.com, pushmeet@google.com, drk@calicolabs.com\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tFnAHhx-ze9X"
      },
      "source": [
        "**Note:** This colab will not yet work since the model isn't yet publicly available. We are working on enabling this and will update the colab accordingly."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "si-w2NPretDg"
      },
      "source": [
        "### Steps\n",
        "\n",
        "This colab demonstrates how to\n",
        "- Make predictions with Enformer and reproduce Fig. 1d\n",
        "- Compute contribution scores and reproduce parts of Fig. 2a\n",
        "- Predict the effect of a genetic variant and reproduce parts of Fig. 3g\n",
        "- Score multiple variants in a VCF "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wCCJsjaHwTYC"
      },
      "source": [
        "## Setup"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NqR7ol3rxrtM"
      },
      "source": [
        "**Start the colab kernel with GPU**: Runtime -> Change runtime type -> GPU"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hTGOLrbZxNHK"
      },
      "source": [
        "import tensorflow as tf\n",
        "# Make sure the GPU is enabled \n",
        "assert tf.config.list_physical_devices('GPU'), 'Start the colab kernel with GPU: Runtime -> Change runtime type -> GPU'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Eg8hcb45wqMM"
      },
      "source": [
        "!pip install kipoiseq --quiet > /dev/null"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MCDk7UQPG0Lr"
      },
      "source": [
        "### Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NRI9KisU11bM"
      },
      "source": [
        "import joblib\n",
        "import gzip\n",
        "import kipoiseq\n",
        "from kipoiseq import Interval\n",
        "import pyfaidx\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib as mpl\n",
        "import seaborn as sns\n",
        "\n",
        "%matplotlib inline\n",
        "%config InlineBackend.figure_format = 'retina'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g0F1A9AaCrkQ"
      },
      "source": [
        "transform_base_path = 'dna_transformer_2020_1kG_SAD.robustscaler-PCA500-robustscaler.transform.pkl'\n",
        "model_base_path = '13376309-3.finetuned'\n",
        "\n",
        "transform_path = f'/root/models/{transform_base_path}'\n",
        "model_path = f'/root/models/{model_base_path}'\n",
        "fasta_file = '/root/data/genome.fa'\n",
        "\n",
        "# Download targets from Basenji2 dataset \n",
        "# Cite: Kelley et al Cross-species regulatory sequence activity prediction. PLoS Comput. Biol. 16, e1008050 (2020).\n",
        "targets_txt = 'https://raw.githubusercontent.com/calico/basenji/master/manuscripts/cross2020/targets_human.txt'\n",
        "clinvar_vcf = '/root/data/clinvar.vcf.gz'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 142
        },
        "id": "OlE6JAVfI08a",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610491573962,
          "user_tz": -60,
          "elapsed": 835,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "65dae93c-91a9-4f89-e8ad-4189a8b71ab7"
      },
      "source": [
        "df_targets = pd.read_csv(targets_txt, sep='\\t')\n",
        "df_targets.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>genome</th>\n",
              "      <th>identifier</th>\n",
              "      <th>file</th>\n",
              "      <th>clip</th>\n",
              "      <th>scale</th>\n",
              "      <th>sum_stat</th>\n",
              "      <th>description</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>ENCFF833POA</td>\n",
              "      <td>/home/drk/tillage/datasets/human/dnase/encode/...</td>\n",
              "      <td>32</td>\n",
              "      <td>2</td>\n",
              "      <td>mean</td>\n",
              "      <td>DNASE:cerebellum male adult (27 years) and mal...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>ENCFF110QGM</td>\n",
              "      <td>/home/drk/tillage/datasets/human/dnase/encode/...</td>\n",
              "      <td>32</td>\n",
              "      <td>2</td>\n",
              "      <td>mean</td>\n",
              "      <td>DNASE:frontal cortex male adult (27 years) and...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>ENCFF880MKD</td>\n",
              "      <td>/home/drk/tillage/datasets/human/dnase/encode/...</td>\n",
              "      <td>32</td>\n",
              "      <td>2</td>\n",
              "      <td>mean</td>\n",
              "      <td>DNASE:chorion</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   index  genome  ... sum_stat                                        description\n",
              "0      0       0  ...     mean  DNASE:cerebellum male adult (27 years) and mal...\n",
              "1      1       0  ...     mean  DNASE:frontal cortex male adult (27 years) and...\n",
              "2      2       0  ...     mean                                      DNASE:chorion\n",
              "\n",
              "[3 rows x 8 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q8ZhswycGux3"
      },
      "source": [
        "### Download files"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_o7RuTMi1kYI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610491604608,
          "user_tz": -60,
          "elapsed": 32,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "cba1dab0-ba63-47d3-cd31-d30acbb2170a"
      },
      "source": [
        "# Mount google drive. Models will be made available through https://www.tensorflow.org/hub\n",
        "# and https://storage.googleapis.com/deepmind-enformer soon.\n",
        "import os.path\n",
        "from google.colab import drive\n",
        "drive.mount('/gdrive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rPk4M0jBCI-k"
      },
      "source": [
        "# Download models\n",
        "!mkdir -p /root/models/\n",
        "models_dir = '/gdrive/My\\ Drive/models/'\n",
        "!cp -R {models_dir}/{transform_base_path} /root/models/\n",
        "!cp -R {models_dir}/{model_base_path} /root/models/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dowTJknFJOHu"
      },
      "source": [
        "Download and index the reference genome fasta file\n",
        "\n",
        "Credit to Genome Reference Consortium: https://www.ncbi.nlm.nih.gov/grc\n",
        "\n",
        "Schneider et al 2017 http://dx.doi.org/10.1101/gr.213611.116: Evaluation of GRCh38 and de novo haploid genome assemblies demonstrates the enduring quality of the reference assembly"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "flOUYxP7Fjvh",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610491763552,
          "user_tz": -60,
          "elapsed": 164448,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "5afcc88a-cf39-4142-e5df-f577a5d6bdd8"
      },
      "source": [
        "\n",
        "!mkdir -p /root/data\n",
        "!wget -O - http://hgdownload.cse.ucsc.edu/goldenPath/hg38/bigZips/hg38.fa.gz | gunzip -c > {fasta_file}\n",
        "pyfaidx.Faidx(fasta_file)\n",
        "!ls /root/data"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2021-01-12 22:47:12--  http://hgdownload.cse.ucsc.edu/goldenPath/hg38/bigZips/hg38.fa.gz\n",
            "Resolving hgdownload.cse.ucsc.edu (hgdownload.cse.ucsc.edu)... 128.114.119.163\n",
            "Connecting to hgdownload.cse.ucsc.edu (hgdownload.cse.ucsc.edu)|128.114.119.163|:80... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 983659424 (938M) [application/x-gzip]\n",
            "Saving to: ‘STDOUT’\n",
            "\n",
            "-                   100%[===================>] 938.09M  10.4MB/s    in 79s     \n",
            "\n",
            "2021-01-12 22:48:31 (11.9 MB/s) - written to stdout [983659424/983659424]\n",
            "\n",
            "genome.fa  genome.fa.fai\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VwMZhF42JH18"
      },
      "source": [
        "Download the clinvar file. Reference:\n",
        "\n",
        "Landrum MJ, Lee JM, Benson M, Brown GR, Chao C, Chitipiralla S, Gu B, Hart J, Hoffman D, Jang W, Karapetyan K, Katz K, Liu C, Maddipatla Z, Malheiro A, McDaniel K, Ovetsky M, Riley G, Zhou G, Holmes JB, Kattman BL, Maglott DR. ClinVar: improving access to variant interpretations and supporting evidence. Nucleic Acids Res . 2018 Jan 4. PubMed PMID: 29165669 .\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4PjXFN5OcGbH",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610491766389,
          "user_tz": -60,
          "elapsed": 167174,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "dd1dafe7-426b-4e73-9c9b-3f2ab6cc7aa6"
      },
      "source": [
        "!wget https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/clinvar.vcf.gz -O /root/data/clinvar.vcf.gz"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2021-01-12 22:49:23--  https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/clinvar.vcf.gz\n",
            "Resolving ftp.ncbi.nlm.nih.gov (ftp.ncbi.nlm.nih.gov)... 130.14.250.7, 2607:f220:41e:250::10, 2607:f220:41e:250::11, ...\n",
            "Connecting to ftp.ncbi.nlm.nih.gov (ftp.ncbi.nlm.nih.gov)|130.14.250.7|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 34090716 (33M) [application/x-gzip]\n",
            "Saving to: ‘/root/data/clinvar.vcf.gz’\n",
            "\n",
            "/root/data/clinvar. 100%[===================>]  32.51M  16.1MB/s    in 2.0s    \n",
            "\n",
            "2021-01-12 22:49:25 (16.1 MB/s) - ‘/root/data/clinvar.vcf.gz’ saved [34090716/34090716]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Omj-KERcwSdB"
      },
      "source": [
        "### Code (double click on the title to show the code)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "cellView": "form",
        "id": "47E4AEgLx1VT"
      },
      "source": [
        "# @title `Enformer`, `EnformerScoreVariantsNormalized`, `EnformerScoreVariantsPCANormalized`,\n",
        "SEQUENCE_LENGTH = 393216\n",
        "\n",
        "class Enformer:\n",
        "\n",
        "  def __init__(self, saved_model_path):\n",
        "    self._model = tf.saved_model.load(saved_model_path).model\n",
        "\n",
        "  def predict_on_batch(self, inputs):\n",
        "    predictions = self._model.predict_on_batch(inputs)\n",
        "    return {k: v.numpy() for k, v in predictions.items()}\n",
        "\n",
        "  @tf.function\n",
        "  def contribution_input_grad(self, input_sequence,\n",
        "                              target_mask, output_head='human'):\n",
        "    input_sequence = input_sequence[tf.newaxis]\n",
        "\n",
        "    target_mask_mass = tf.reduce_sum(target_mask)\n",
        "    with tf.GradientTape() as tape:\n",
        "      tape.watch(input_sequence)\n",
        "      prediction = tf.reduce_sum(\n",
        "          target_mask[tf.newaxis] *\n",
        "          self._model.predict_on_batch(input_sequence)[output_head]) / target_mask_mass\n",
        "\n",
        "    input_grad = tape.gradient(prediction, input_sequence) * input_sequence\n",
        "    input_grad = tf.squeeze(input_grad, axis=0)\n",
        "    return tf.reduce_sum(input_grad, axis=-1)\n",
        "\n",
        "\n",
        "class EnformerScoreVariantsRaw:\n",
        "\n",
        "  def __init__(self, saved_model_path, organism='human'):\n",
        "    self._model = Enformer(saved_model_path)\n",
        "    self._organism = organism\n",
        "  \n",
        "  def predict_on_batch(self, inputs):\n",
        "    ref_prediction = self._model.predict_on_batch(inputs['ref'])[self._organism]\n",
        "    alt_prediction = self._model.predict_on_batch(inputs['alt'])[self._organism]\n",
        "\n",
        "    return alt_prediction.mean(axis=1) - ref_prediction.mean(axis=1)\n",
        "\n",
        "\n",
        "class EnformerScoreVariantsNormalized:\n",
        "\n",
        "  def __init__(self, saved_model_path, transform_pkl_path,\n",
        "               organism='human'):\n",
        "    assert organism == 'human', 'Transforms only compatible with organism=human'\n",
        "    self._model = EnformerScoreVariantsRaw(saved_model_path, organism)\n",
        "    transform_pipeline = joblib.load(transform_pkl_path)\n",
        "    self._transform = transform_pipeline.steps[0][1]  # StandardScaler.\n",
        "    \n",
        "  def predict_on_batch(self, inputs):\n",
        "    scores = self._model.predict_on_batch(inputs)\n",
        "    return self._transform.transform(scores)\n",
        "\n",
        "\n",
        "class EnformerScoreVariantsPCANormalized:\n",
        "\n",
        "  def __init__(self, saved_model_path, transform_pkl_path,\n",
        "               organism='human', num_top_features=500):\n",
        "    self._model = EnformerScoreVariantsRaw(saved_model_path, organism)\n",
        "    self._transform = joblib.load(transform_pkl_path)\n",
        "    self._num_top_features = num_top_features\n",
        "    \n",
        "  def predict_on_batch(self, inputs):\n",
        "    scores = self._model.predict_on_batch(inputs)\n",
        "    return self._transform.transform(scores)[:, :self._num_top_features]\n",
        "\n",
        "\n",
        "# TODO(avsec): Add feature description: Either PCX, or full names."
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "cellView": "form",
        "id": "uLYRbOXDyA36"
      },
      "source": [
        "# @title `variant_centered_sequences`\n",
        "\n",
        "class FastaStringExtractor:\n",
        "    \n",
        "    def __init__(self, fasta_file):\n",
        "        self.fasta = pyfaidx.Fasta(fasta_file)\n",
        "        self._chromosome_sizes = {k: len(v) for k, v in self.fasta.items()}\n",
        "\n",
        "    def extract(self, interval: Interval, **kwargs) -> str:\n",
        "        # Truncate interval if it extends beyond the chromosome lengths.\n",
        "        chromosome_length = self._chromosome_sizes[interval.chrom]\n",
        "        trimmed_interval = Interval(interval.chrom,\n",
        "                                    max(interval.start, 0),\n",
        "                                    min(interval.end, chromosome_length),\n",
        "                                    )\n",
        "        # pyfaidx wants a 1-based interval\n",
        "        sequence = str(self.fasta.get_seq(trimmed_interval.chrom,\n",
        "                                          trimmed_interval.start + 1,\n",
        "                                          trimmed_interval.stop).seq).upper()\n",
        "        # Fill truncated values with N's.\n",
        "        pad_upstream = 'N' * max(-interval.start, 0)\n",
        "        pad_downstream = 'N' * max(interval.end - chromosome_length, 0)\n",
        "        return pad_upstream + sequence + pad_downstream\n",
        "\n",
        "    def close(self):\n",
        "        return self.fasta.close()\n",
        "\n",
        "\n",
        "def variant_generator(vcf_file, gzipped=False):\n",
        "  \"\"\"Yields a kipoiseq.dataclasses.Variant for each row in VCF file.\"\"\"\n",
        "  def _open(file):\n",
        "    return gzip.open(vcf_file, 'rt') if gzipped else open(vcf_file)\n",
        "    \n",
        "  with _open(vcf_file) as f:\n",
        "    for line in f:\n",
        "      if line.startswith('#'):\n",
        "        continue\n",
        "      chrom, pos, id, ref, alt_list = line.split('\\t')[:5]\n",
        "      # Split ALT alleles and return individual variants as output.\n",
        "      for alt in alt_list.split(','):\n",
        "        yield kipoiseq.dataclasses.Variant(chrom=chrom, pos=pos,\n",
        "                                           ref=ref, alt=alt, id=id)\n",
        "\n",
        "\n",
        "def one_hot_encode(sequence):\n",
        "  return kipoiseq.transforms.functional.one_hot_dna(sequence).astype(np.float32)\n",
        "\n",
        "\n",
        "def variant_centered_sequences(vcf_file, sequence_length, gzipped=False,\n",
        "                               chr_prefix=''):\n",
        "  seq_extractor = kipoiseq.extractors.VariantSeqExtractor(\n",
        "    reference_sequence=FastaStringExtractor(fasta_file))\n",
        "\n",
        "  for variant in variant_generator(vcf_file, gzipped=gzipped):\n",
        "    interval = Interval(chr_prefix + variant.chrom,\n",
        "                        variant.pos, variant.pos)\n",
        "    interval = interval.resize(sequence_length)\n",
        "    center = interval.center() - interval.start\n",
        "\n",
        "    reference = seq_extractor.extract(interval, [], anchor=center)\n",
        "    alternate = seq_extractor.extract(interval, [variant], anchor=center)\n",
        "\n",
        "    yield {'inputs': {'ref': one_hot_encode(reference),\n",
        "                      'alt': one_hot_encode(alternate)},\n",
        "           'metadata': {'chrom': chr_prefix + variant.chrom,\n",
        "                        'pos': variant.pos,\n",
        "                        'id': variant.id,\n",
        "                        'ref': variant.ref,\n",
        "                        'alt': variant.alt}}"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "cellView": "form",
        "id": "Up1oCMFiPucp"
      },
      "source": [
        "# @title `plot_tracks`\n",
        "\n",
        "def plot_tracks(tracks, interval, height=1.5):\n",
        "  fig, axes = plt.subplots(len(tracks), 1, figsize=(20, height * len(tracks)), sharex=True)\n",
        "  for ax, (title, y) in zip(axes, tracks.items()):\n",
        "    ax.fill_between(np.linspace(interval.start, interval.end, num=len(y)), y)\n",
        "    ax.set_title(title)\n",
        "    sns.despine(top=True, right=True, bottom=True)\n",
        "  ax.set_xlabel(str(interval))\n",
        "  plt.tight_layout()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zEwfoz3cwOzt"
      },
      "source": [
        "## Make predictions for a genetic sequenece"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WC-pgC35DgnL"
      },
      "source": [
        "model = Enformer(model_path)\n",
        "\n",
        "fasta_extractor = FastaStringExtractor(fasta_file)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8u8Gt8WWyG53"
      },
      "source": [
        "# @title Make predictions for an genomic example interval\n",
        "target_interval = kipoiseq.Interval('chr11', 35_082_742, 35_197_430)  # @param\n",
        "\n",
        "sequence_one_hot = one_hot_encode(fasta_extractor.extract(target_interval.resize(SEQUENCE_LENGTH)))\n",
        "predictions = model.predict_on_batch(sequence_one_hot[np.newaxis])['human'][0]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 441
        },
        "id": "x7RrBbPFJdf8",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610403486072,
          "user_tz": -60,
          "elapsed": 29386,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "c25e94f8-ee03-4b4d-8f3e-b9610a0f4527"
      },
      "source": [
        "# @title Plot tracks\n",
        "tracks = {'DNASE:CD14-positive monocyte female': predictions[:, 41],\n",
        "          'DNASE:keratinocyte female': predictions[:, 42],\n",
        "          'CHIP:H3K27ac:keratinocyte female': predictions[:, 706],\n",
        "          'CAGE:Keratinocyte - epidermal': np.log10(1 + predictions[:, 4799])}\n",
        "plot_tracks(tracks, target_interval)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x432 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gM2KwV8gwMNj"
      },
      "source": [
        "## Contribution scores example"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o4c2W_MjNBzp"
      },
      "source": [
        "# @title Compute contribution scores\n",
        "target_interval = kipoiseq.Interval('chr12', 54_223_589, 54_338_277)  # @param\n",
        "\n",
        "sequence_one_hot = one_hot_encode(fasta_extractor.extract(target_interval.resize(SEQUENCE_LENGTH)))\n",
        "predictions = model.predict_on_batch(sequence_one_hot[np.newaxis])['human'][0]\n",
        "\n",
        "target_mask = np.zeros_like(predictions)\n",
        "for idx in [447, 448, 449]:\n",
        "  target_mask[idx, 4828] = 1\n",
        "  target_mask[idx, 5111] = 1\n",
        "# This will take some time since tf.function needs to get compiled.\n",
        "contribution_scores = model.contribution_input_grad(sequence_one_hot.astype(np.float32), target_mask).numpy()\n",
        "pooled_contribution_scores = tf.nn.avg_pool1d(np.abs(contribution_scores)[np.newaxis, :, np.newaxis], 128, 128, 'VALID')[0, :, 0].numpy()[1088:-1088]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "id": "nzHZiE40Wnyk",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610403508617,
          "user_tz": -60,
          "elapsed": 51920,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "68439c7d-1e26-417f-b6f2-a858e3c9fece"
      },
      "source": [
        "tracks = {'CAGE predictions': predictions[:, 4828],\n",
        "          'Enformer gradient*input': np.minimum(pooled_contribution_scores, 0.03)}\n",
        "plot_tracks(tracks, target_interval);"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x216 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iiDN_ScSv3sI"
      },
      "source": [
        "## Variant scoring example"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hIaN9GKDVwLK"
      },
      "source": [
        "# @title Score the variant\n",
        "variant = kipoiseq.Variant('chr16', 57025062, 'C', 'T', id='rs11644125')  # @param\n",
        "\n",
        "# Center the interval at the variant\n",
        "interval = kipoiseq.Interval(variant.chrom, variant.start, variant.start).resize(SEQUENCE_LENGTH)\n",
        "seq_extractor = kipoiseq.extractors.VariantSeqExtractor(reference_sequence=fasta_extractor)\n",
        "center = interval.center() - interval.start\n",
        "\n",
        "reference = seq_extractor.extract(interval, [], anchor=center)\n",
        "alternate = seq_extractor.extract(interval, [variant], anchor=center)\n",
        "\n",
        "# Make predictions for the refernece and alternate allele\n",
        "reference_prediction = model.predict_on_batch(one_hot_encode(reference)[np.newaxis])['human'][0]\n",
        "alternate_prediction = model.predict_on_batch(one_hot_encode(alternate)[np.newaxis])['human'][0]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 368
        },
        "id": "Bczbku-FUoGr",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610491873418,
          "user_tz": -60,
          "elapsed": 1993,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "95294fac-824d-4a82-864d-2b246b7d7973"
      },
      "source": [
        "# @title Visualize some tracks\n",
        "variant_track = np.zeros_like(reference_prediction[:, 0], dtype=bool)\n",
        "variant_track[variant_track.shape[0] // 2] = True\n",
        "tracks = {'variant': variant_track,\n",
        "          'CAGE/neutrofils ref': reference_prediction[:, 4767],\n",
        "          'CAGE/neutrofils alt-ref': alternate_prediction[:, 4767] - reference_prediction[:, 4767],\n",
        "          'CHIP:H3K27ac:neutrophil ref': reference_prediction[:, 2280],\n",
        "          'CHIP:H3K27ac:neutrophil alt-ref': alternate_prediction[:, 2280] - reference_prediction[:, 2280],\n",
        "          }\n",
        "\n",
        "plot_tracks(tracks, interval.resize(reference_prediction.shape[0] * 128), height=1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x360 with 5 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 1436,
              "height": 351
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZMCU9woQv6Ea"
      },
      "source": [
        "## Score variants in a VCF file"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c1nJwFS-4P8g"
      },
      "source": [
        "### Report top 20 PCs"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-vmj1MA3chRQ"
      },
      "source": [
        "enformer_score_variants = EnformerScoreVariantsPCANormalized(model_path, transform_path, num_top_features=20)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 275
        },
        "id": "1oGEYix3dRex",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610403541225,
          "user_tz": -60,
          "elapsed": 84507,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "85c13ade-4e8c-42af-ec36-984f847fe513"
      },
      "source": [
        "# Score the first 5 variants from ClinVar\n",
        "# Lower-dimensional scores (20 PCs)\n",
        "it = variant_centered_sequences(clinvar_vcf, sequence_length=SEQUENCE_LENGTH,\n",
        "                                 gzipped=True, chr_prefix='chr')\n",
        "example_list = []\n",
        "for i, example in enumerate(it):\n",
        "  if i >= 5:\n",
        "    break\n",
        "  variant_scores = enformer_score_variants.predict_on_batch(\n",
        "      {k: v[tf.newaxis] for k,v in example['inputs'].items()})[0]\n",
        "  variant_scores = {f'PC{i}': score for i, score in enumerate(variant_scores)}\n",
        "  example_list.append({**example['metadata'],\n",
        "                       **variant_scores})\n",
        "  if i % 2 == 0:\n",
        "    print(f'Done {i}')\n",
        "df = pd.DataFrame(example_list)\n",
        "df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Done 0\n",
            "Done 2\n",
            "Done 4\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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>chrom</th>\n",
              "      <th>pos</th>\n",
              "      <th>id</th>\n",
              "      <th>ref</th>\n",
              "      <th>alt</th>\n",
              "      <th>PC0</th>\n",
              "      <th>PC1</th>\n",
              "      <th>PC2</th>\n",
              "      <th>PC3</th>\n",
              "      <th>PC4</th>\n",
              "      <th>PC5</th>\n",
              "      <th>PC6</th>\n",
              "      <th>PC7</th>\n",
              "      <th>PC8</th>\n",
              "      <th>PC9</th>\n",
              "      <th>PC10</th>\n",
              "      <th>PC11</th>\n",
              "      <th>PC12</th>\n",
              "      <th>PC13</th>\n",
              "      <th>PC14</th>\n",
              "      <th>PC15</th>\n",
              "      <th>PC16</th>\n",
              "      <th>PC17</th>\n",
              "      <th>PC18</th>\n",
              "      <th>PC19</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930188</td>\n",
              "      <td>846933</td>\n",
              "      <td>G</td>\n",
              "      <td>A</td>\n",
              "      <td>-61.471287</td>\n",
              "      <td>-5.655146</td>\n",
              "      <td>-2.758397</td>\n",
              "      <td>6.290497</td>\n",
              "      <td>1.846261</td>\n",
              "      <td>3.449631</td>\n",
              "      <td>6.052878</td>\n",
              "      <td>0.634326</td>\n",
              "      <td>2.584449</td>\n",
              "      <td>1.118316</td>\n",
              "      <td>1.944067</td>\n",
              "      <td>-6.512556</td>\n",
              "      <td>-4.948595</td>\n",
              "      <td>-1.168765</td>\n",
              "      <td>-2.903484</td>\n",
              "      <td>0.484869</td>\n",
              "      <td>2.897006</td>\n",
              "      <td>1.758686</td>\n",
              "      <td>3.685951</td>\n",
              "      <td>-6.672137</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930203</td>\n",
              "      <td>972363</td>\n",
              "      <td>C</td>\n",
              "      <td>T</td>\n",
              "      <td>21.482735</td>\n",
              "      <td>-9.249768</td>\n",
              "      <td>3.685621</td>\n",
              "      <td>9.606244</td>\n",
              "      <td>0.538989</td>\n",
              "      <td>5.262571</td>\n",
              "      <td>1.418416</td>\n",
              "      <td>-14.005647</td>\n",
              "      <td>12.897153</td>\n",
              "      <td>9.390783</td>\n",
              "      <td>-5.201140</td>\n",
              "      <td>-3.090504</td>\n",
              "      <td>1.975024</td>\n",
              "      <td>-5.241168</td>\n",
              "      <td>-14.366508</td>\n",
              "      <td>-7.802508</td>\n",
              "      <td>0.141153</td>\n",
              "      <td>12.087668</td>\n",
              "      <td>-5.559410</td>\n",
              "      <td>9.170748</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930248</td>\n",
              "      <td>789256</td>\n",
              "      <td>G</td>\n",
              "      <td>A</td>\n",
              "      <td>0.498874</td>\n",
              "      <td>1.982327</td>\n",
              "      <td>-2.750614</td>\n",
              "      <td>-3.141907</td>\n",
              "      <td>-5.241604</td>\n",
              "      <td>-4.961285</td>\n",
              "      <td>-1.702641</td>\n",
              "      <td>2.543528</td>\n",
              "      <td>-2.248600</td>\n",
              "      <td>-0.774210</td>\n",
              "      <td>-0.368460</td>\n",
              "      <td>-1.392373</td>\n",
              "      <td>-6.709529</td>\n",
              "      <td>0.173268</td>\n",
              "      <td>6.361717</td>\n",
              "      <td>3.989930</td>\n",
              "      <td>1.602195</td>\n",
              "      <td>-5.265455</td>\n",
              "      <td>0.852734</td>\n",
              "      <td>-0.946416</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930275</td>\n",
              "      <td>969662</td>\n",
              "      <td>T</td>\n",
              "      <td>G</td>\n",
              "      <td>-3.028966</td>\n",
              "      <td>2.984429</td>\n",
              "      <td>7.532916</td>\n",
              "      <td>6.588795</td>\n",
              "      <td>19.307785</td>\n",
              "      <td>5.661951</td>\n",
              "      <td>0.962246</td>\n",
              "      <td>5.863384</td>\n",
              "      <td>-6.041630</td>\n",
              "      <td>2.309155</td>\n",
              "      <td>9.062576</td>\n",
              "      <td>1.426052</td>\n",
              "      <td>-3.792181</td>\n",
              "      <td>3.136441</td>\n",
              "      <td>-2.617620</td>\n",
              "      <td>0.942220</td>\n",
              "      <td>-6.195234</td>\n",
              "      <td>-6.291888</td>\n",
              "      <td>-5.329431</td>\n",
              "      <td>-1.141422</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930336</td>\n",
              "      <td>843786</td>\n",
              "      <td>G</td>\n",
              "      <td>A</td>\n",
              "      <td>-45.674007</td>\n",
              "      <td>-6.556405</td>\n",
              "      <td>-0.550849</td>\n",
              "      <td>16.986986</td>\n",
              "      <td>-5.257315</td>\n",
              "      <td>5.071064</td>\n",
              "      <td>6.669114</td>\n",
              "      <td>-1.328647</td>\n",
              "      <td>-9.566864</td>\n",
              "      <td>11.220359</td>\n",
              "      <td>3.839488</td>\n",
              "      <td>-8.214136</td>\n",
              "      <td>-4.864720</td>\n",
              "      <td>-2.947721</td>\n",
              "      <td>-13.659990</td>\n",
              "      <td>1.405919</td>\n",
              "      <td>-3.936957</td>\n",
              "      <td>10.316927</td>\n",
              "      <td>1.663230</td>\n",
              "      <td>-5.247654</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "  chrom     pos      id ref  ...      PC16       PC17      PC18      PC19\n",
              "0  chr1  930188  846933   G  ...  2.897006   1.758686  3.685951 -6.672137\n",
              "1  chr1  930203  972363   C  ...  0.141153  12.087668 -5.559410  9.170748\n",
              "2  chr1  930248  789256   G  ...  1.602195  -5.265455  0.852734 -0.946416\n",
              "3  chr1  930275  969662   T  ... -6.195234  -6.291888 -5.329431 -1.141422\n",
              "4  chr1  930336  843786   G  ... -3.936957  10.316927  1.663230 -5.247654\n",
              "\n",
              "[5 rows x 25 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_G5cANX34SLw"
      },
      "source": [
        "### Report all 5,313 features (z-score normalized)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "get8hogCySnt"
      },
      "source": [
        "enformer_score_variants_all = EnformerScoreVariantsNormalized(model_path, transform_path)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 338
        },
        "id": "Q48earqRyFa6",
        "executionInfo": {
          "status": "ok",
          "timestamp": 1610491940923,
          "user_tz": -60,
          "elapsed": 22163,
          "user": {
            "displayName": "Žiga Avsec",
            "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GjeqpTSqpdiH6-w18IlBq5jUYLVI3qWge6m6wC5=s64",
            "userId": "06962645976249714923"
          }
        },
        "outputId": "dd7629d6-fdcb-4057-d72a-5c75cdc9b3b6"
      },
      "source": [
        "# Score the first 5 variants from ClinVar\n",
        "# All Scores\n",
        "it = variant_centered_sequences(clinvar_vcf, sequence_length=SEQUENCE_LENGTH,\n",
        "                                 gzipped=True, chr_prefix='chr')\n",
        "example_list = []\n",
        "for i, example in enumerate(it):\n",
        "  if i >= 5:\n",
        "    break\n",
        "  variant_scores = enformer_score_variants_all.predict_on_batch(\n",
        "      {k: v[tf.newaxis] for k,v in example['inputs'].items()})[0]\n",
        "  variant_scores = {name[:20]: score for name, score in zip(df_targets.description, variant_scores)}\n",
        "  example_list.append({**example['metadata'],\n",
        "                       **variant_scores})\n",
        "  if i % 2 == 0:\n",
        "    print(f'Done {i}')\n",
        "df = pd.DataFrame(example_list)\n",
        "df"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Done 0\n",
            "Done 2\n",
            "Done 4\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "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>chrom</th>\n",
              "      <th>pos</th>\n",
              "      <th>id</th>\n",
              "      <th>ref</th>\n",
              "      <th>alt</th>\n",
              "      <th>DNASE:cerebellum mal</th>\n",
              "      <th>DNASE:frontal cortex</th>\n",
              "      <th>DNASE:chorion</th>\n",
              "      <th>DNASE:Ishikawa treat</th>\n",
              "      <th>DNASE:GM03348</th>\n",
              "      <th>DNASE:GM03348 geneti</th>\n",
              "      <th>DNASE:AG08395</th>\n",
              "      <th>DNASE:AG08396</th>\n",
              "      <th>DNASE:AG20443</th>\n",
              "      <th>DNASE:H54</th>\n",
              "      <th>DNASE:GM10248</th>\n",
              "      <th>DNASE:GM12878</th>\n",
              "      <th>DNASE:GM12891</th>\n",
              "      <th>DNASE:GM12892</th>\n",
              "      <th>DNASE:GM18507</th>\n",
              "      <th>DNASE:GM19238</th>\n",
              "      <th>DNASE:GM19239</th>\n",
              "      <th>DNASE:GM19240</th>\n",
              "      <th>DNASE:H1-hESC</th>\n",
              "      <th>DNASE:H7-hESC</th>\n",
              "      <th>DNASE:H9</th>\n",
              "      <th>DNASE:heart male adu</th>\n",
              "      <th>DNASE:HEK293T</th>\n",
              "      <th>DNASE:HeLa-S3 treate</th>\n",
              "      <th>DNASE:HeLa-S3</th>\n",
              "      <th>DNASE:hepatocyte</th>\n",
              "      <th>DNASE:HepG2</th>\n",
              "      <th>DNASE:HTR-8/SVneo</th>\n",
              "      <th>DNASE:endothelial ce</th>\n",
              "      <th>DNASE:CWRU1 male</th>\n",
              "      <th>DNASE:iPS-NIHi11 mal</th>\n",
              "      <th>DNASE:iPS-NIHi7 fema</th>\n",
              "      <th>DNASE:K562 treated w</th>\n",
              "      <th>DNASE:K562 G2 phase</th>\n",
              "      <th>DNASE:K562 G1 phase</th>\n",
              "      <th>...</th>\n",
              "      <th>CAGE:CD14+CD16- Mono</th>\n",
              "      <th>CAGE:achilles tendon</th>\n",
              "      <th>CAGE:cerebrospinal f</th>\n",
              "      <th>CAGE:cruciate ligame</th>\n",
              "      <th>CAGE:eye - vitreous</th>\n",
              "      <th>CAGE:eye - muscle su</th>\n",
              "      <th>CAGE:eye - muscle la</th>\n",
              "      <th>CAGE:eye - muscle me</th>\n",
              "      <th>CAGE:eye - muscle in</th>\n",
              "      <th>CAGE:Fingernail (inc</th>\n",
              "      <th>CAGE:optic nerve,</th>\n",
              "      <th>CAGE:Skin - palm,</th>\n",
              "      <th>CAGE:tongue epidermi</th>\n",
              "      <th>CAGE:Urethra,</th>\n",
              "      <th>CAGE:CD14+ monocytes</th>\n",
              "      <th>CAGE:Hep-2 cells tre</th>\n",
              "      <th>CAGE:Hep-2 cells moc</th>\n",
              "      <th>CAGE:immature langer</th>\n",
              "      <th>CAGE:migratory lange</th>\n",
              "      <th>CAGE:CD34 cells diff</th>\n",
              "      <th>CAGE:amygdala - adul</th>\n",
              "      <th>CAGE:thalamus - adul</th>\n",
              "      <th>CAGE:hippocampus - a</th>\n",
              "      <th>CAGE:parietal lobe -</th>\n",
              "      <th>CAGE:cerebellum - ad</th>\n",
              "      <th>CAGE:pineal gland -</th>\n",
              "      <th>CAGE:spinal cord - a</th>\n",
              "      <th>CAGE:Olfactory epith</th>\n",
              "      <th>CAGE:gamma delta pos</th>\n",
              "      <th>CAGE:Mast cell, expa</th>\n",
              "      <th>CAGE:adipose,</th>\n",
              "      <th>CAGE:cerebellum, new</th>\n",
              "      <th>CAGE:spinal cord, ne</th>\n",
              "      <th>CAGE:amygdala, newbo</th>\n",
              "      <th>CAGE:hippocampus, ne</th>\n",
              "      <th>CAGE:putamen, newbor</th>\n",
              "      <th>CAGE:thalamus, newbo</th>\n",
              "      <th>CAGE:thymic carcinom</th>\n",
              "      <th>CAGE:Smooth muscle c</th>\n",
              "      <th>CAGE:parietal cortex</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930188</td>\n",
              "      <td>846933</td>\n",
              "      <td>G</td>\n",
              "      <td>A</td>\n",
              "      <td>-1.786843</td>\n",
              "      <td>-15.682123</td>\n",
              "      <td>-22.520702</td>\n",
              "      <td>0.528606</td>\n",
              "      <td>-3.829948</td>\n",
              "      <td>-3.232739</td>\n",
              "      <td>-2.671064</td>\n",
              "      <td>-9.467652</td>\n",
              "      <td>-3.155200</td>\n",
              "      <td>-7.706491</td>\n",
              "      <td>-3.149441</td>\n",
              "      <td>-0.523933</td>\n",
              "      <td>-11.134032</td>\n",
              "      <td>-11.109153</td>\n",
              "      <td>-3.670341</td>\n",
              "      <td>-8.464477</td>\n",
              "      <td>-11.757355</td>\n",
              "      <td>-8.598063</td>\n",
              "      <td>-0.531494</td>\n",
              "      <td>0.225998</td>\n",
              "      <td>-0.777080</td>\n",
              "      <td>-15.741483</td>\n",
              "      <td>0.879350</td>\n",
              "      <td>-6.576440</td>\n",
              "      <td>-4.204203</td>\n",
              "      <td>-8.632437</td>\n",
              "      <td>-0.466419</td>\n",
              "      <td>-3.439260</td>\n",
              "      <td>-0.311319</td>\n",
              "      <td>-4.804222</td>\n",
              "      <td>-5.151290</td>\n",
              "      <td>-1.418189</td>\n",
              "      <td>0.366935</td>\n",
              "      <td>-2.001119</td>\n",
              "      <td>-2.294838</td>\n",
              "      <td>...</td>\n",
              "      <td>-33.057209</td>\n",
              "      <td>-2.997374</td>\n",
              "      <td>-8.419192</td>\n",
              "      <td>-5.191596</td>\n",
              "      <td>-4.536147</td>\n",
              "      <td>-5.455176</td>\n",
              "      <td>-3.037405</td>\n",
              "      <td>-7.164758</td>\n",
              "      <td>-2.292249</td>\n",
              "      <td>-8.986766</td>\n",
              "      <td>-4.790454</td>\n",
              "      <td>-4.795162</td>\n",
              "      <td>-1.450022</td>\n",
              "      <td>-1.989321</td>\n",
              "      <td>-22.918665</td>\n",
              "      <td>-35.752350</td>\n",
              "      <td>-35.278088</td>\n",
              "      <td>-13.375848</td>\n",
              "      <td>-28.361494</td>\n",
              "      <td>-31.598894</td>\n",
              "      <td>-11.373973</td>\n",
              "      <td>-9.255016</td>\n",
              "      <td>-9.010527</td>\n",
              "      <td>-9.231528</td>\n",
              "      <td>-3.780572</td>\n",
              "      <td>-0.282759</td>\n",
              "      <td>-12.095373</td>\n",
              "      <td>-59.163910</td>\n",
              "      <td>-66.295670</td>\n",
              "      <td>-34.482758</td>\n",
              "      <td>-21.271309</td>\n",
              "      <td>-4.326236</td>\n",
              "      <td>-16.074772</td>\n",
              "      <td>-6.284147</td>\n",
              "      <td>-10.391891</td>\n",
              "      <td>-6.422609</td>\n",
              "      <td>-9.180881</td>\n",
              "      <td>-14.222376</td>\n",
              "      <td>-46.489288</td>\n",
              "      <td>-11.498157</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930203</td>\n",
              "      <td>972363</td>\n",
              "      <td>C</td>\n",
              "      <td>T</td>\n",
              "      <td>3.888234</td>\n",
              "      <td>-0.792556</td>\n",
              "      <td>-1.675466</td>\n",
              "      <td>4.746172</td>\n",
              "      <td>1.695362</td>\n",
              "      <td>2.216201</td>\n",
              "      <td>-0.281044</td>\n",
              "      <td>-1.136617</td>\n",
              "      <td>0.372290</td>\n",
              "      <td>-3.771606</td>\n",
              "      <td>3.466754</td>\n",
              "      <td>1.097348</td>\n",
              "      <td>2.531235</td>\n",
              "      <td>5.103032</td>\n",
              "      <td>1.873044</td>\n",
              "      <td>3.959806</td>\n",
              "      <td>1.216984</td>\n",
              "      <td>3.810981</td>\n",
              "      <td>0.649070</td>\n",
              "      <td>2.553843</td>\n",
              "      <td>0.989534</td>\n",
              "      <td>1.633624</td>\n",
              "      <td>5.512297</td>\n",
              "      <td>-6.207155</td>\n",
              "      <td>-13.487192</td>\n",
              "      <td>6.362380</td>\n",
              "      <td>0.166047</td>\n",
              "      <td>-4.554312</td>\n",
              "      <td>-0.160736</td>\n",
              "      <td>-1.996482</td>\n",
              "      <td>-2.641517</td>\n",
              "      <td>0.414536</td>\n",
              "      <td>7.214015</td>\n",
              "      <td>4.067241</td>\n",
              "      <td>2.957357</td>\n",
              "      <td>...</td>\n",
              "      <td>33.842506</td>\n",
              "      <td>-0.583508</td>\n",
              "      <td>0.268065</td>\n",
              "      <td>-3.203938</td>\n",
              "      <td>5.382962</td>\n",
              "      <td>1.292423</td>\n",
              "      <td>0.993049</td>\n",
              "      <td>2.095632</td>\n",
              "      <td>-0.293384</td>\n",
              "      <td>4.008999</td>\n",
              "      <td>-0.025227</td>\n",
              "      <td>-0.550521</td>\n",
              "      <td>-1.000858</td>\n",
              "      <td>0.568729</td>\n",
              "      <td>27.443323</td>\n",
              "      <td>-22.869347</td>\n",
              "      <td>-32.034206</td>\n",
              "      <td>12.032639</td>\n",
              "      <td>15.752299</td>\n",
              "      <td>19.407349</td>\n",
              "      <td>2.799562</td>\n",
              "      <td>3.406885</td>\n",
              "      <td>1.643052</td>\n",
              "      <td>2.385034</td>\n",
              "      <td>-1.272377</td>\n",
              "      <td>35.168621</td>\n",
              "      <td>-0.531365</td>\n",
              "      <td>-15.452000</td>\n",
              "      <td>42.711445</td>\n",
              "      <td>22.988682</td>\n",
              "      <td>-2.011288</td>\n",
              "      <td>1.716742</td>\n",
              "      <td>1.979876</td>\n",
              "      <td>5.205636</td>\n",
              "      <td>6.388551</td>\n",
              "      <td>3.159813</td>\n",
              "      <td>0.636698</td>\n",
              "      <td>-0.784998</td>\n",
              "      <td>-4.407548</td>\n",
              "      <td>3.055573</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930248</td>\n",
              "      <td>789256</td>\n",
              "      <td>G</td>\n",
              "      <td>A</td>\n",
              "      <td>-3.392124</td>\n",
              "      <td>-6.586333</td>\n",
              "      <td>-6.679351</td>\n",
              "      <td>-0.481332</td>\n",
              "      <td>0.732646</td>\n",
              "      <td>1.129496</td>\n",
              "      <td>1.167766</td>\n",
              "      <td>-1.667323</td>\n",
              "      <td>0.800817</td>\n",
              "      <td>-4.803396</td>\n",
              "      <td>-0.813693</td>\n",
              "      <td>0.774903</td>\n",
              "      <td>-2.916472</td>\n",
              "      <td>-3.757783</td>\n",
              "      <td>0.205903</td>\n",
              "      <td>-2.631254</td>\n",
              "      <td>-2.177644</td>\n",
              "      <td>-5.314756</td>\n",
              "      <td>-1.064787</td>\n",
              "      <td>-4.775336</td>\n",
              "      <td>-0.307169</td>\n",
              "      <td>-9.369971</td>\n",
              "      <td>-4.574942</td>\n",
              "      <td>-1.763874</td>\n",
              "      <td>4.670806</td>\n",
              "      <td>-9.955566</td>\n",
              "      <td>2.540043</td>\n",
              "      <td>-0.725449</td>\n",
              "      <td>-0.909228</td>\n",
              "      <td>-2.151628</td>\n",
              "      <td>-1.621970</td>\n",
              "      <td>0.037685</td>\n",
              "      <td>-6.617916</td>\n",
              "      <td>-4.519673</td>\n",
              "      <td>-3.751483</td>\n",
              "      <td>...</td>\n",
              "      <td>-3.718003</td>\n",
              "      <td>-0.702731</td>\n",
              "      <td>3.008476</td>\n",
              "      <td>0.761387</td>\n",
              "      <td>-3.265729</td>\n",
              "      <td>1.070971</td>\n",
              "      <td>-0.497848</td>\n",
              "      <td>0.965013</td>\n",
              "      <td>0.367486</td>\n",
              "      <td>-1.758073</td>\n",
              "      <td>0.375682</td>\n",
              "      <td>-0.949173</td>\n",
              "      <td>-1.022737</td>\n",
              "      <td>0.486035</td>\n",
              "      <td>-0.185559</td>\n",
              "      <td>56.499584</td>\n",
              "      <td>67.082275</td>\n",
              "      <td>1.473261</td>\n",
              "      <td>-0.636782</td>\n",
              "      <td>-0.981452</td>\n",
              "      <td>-0.348060</td>\n",
              "      <td>0.946137</td>\n",
              "      <td>0.874985</td>\n",
              "      <td>0.469065</td>\n",
              "      <td>-3.461762</td>\n",
              "      <td>-21.804483</td>\n",
              "      <td>1.403917</td>\n",
              "      <td>-1.323191</td>\n",
              "      <td>0.396306</td>\n",
              "      <td>-5.381416</td>\n",
              "      <td>4.841636</td>\n",
              "      <td>0.019535</td>\n",
              "      <td>2.721695</td>\n",
              "      <td>-1.743816</td>\n",
              "      <td>-1.392196</td>\n",
              "      <td>1.243525</td>\n",
              "      <td>-1.840826</td>\n",
              "      <td>-0.622136</td>\n",
              "      <td>6.409196</td>\n",
              "      <td>-0.923778</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930275</td>\n",
              "      <td>969662</td>\n",
              "      <td>T</td>\n",
              "      <td>G</td>\n",
              "      <td>3.287432</td>\n",
              "      <td>-1.920073</td>\n",
              "      <td>5.813212</td>\n",
              "      <td>2.442116</td>\n",
              "      <td>3.727493</td>\n",
              "      <td>3.496869</td>\n",
              "      <td>0.895325</td>\n",
              "      <td>2.030542</td>\n",
              "      <td>4.058121</td>\n",
              "      <td>-1.889122</td>\n",
              "      <td>1.946017</td>\n",
              "      <td>-1.543661</td>\n",
              "      <td>9.562781</td>\n",
              "      <td>8.657282</td>\n",
              "      <td>1.719417</td>\n",
              "      <td>1.913109</td>\n",
              "      <td>7.135823</td>\n",
              "      <td>3.970765</td>\n",
              "      <td>0.610678</td>\n",
              "      <td>3.715666</td>\n",
              "      <td>0.340136</td>\n",
              "      <td>-12.232494</td>\n",
              "      <td>5.522506</td>\n",
              "      <td>7.254625</td>\n",
              "      <td>7.424074</td>\n",
              "      <td>-1.945692</td>\n",
              "      <td>1.009203</td>\n",
              "      <td>-0.339282</td>\n",
              "      <td>-0.795034</td>\n",
              "      <td>4.926074</td>\n",
              "      <td>9.191165</td>\n",
              "      <td>2.844069</td>\n",
              "      <td>5.490332</td>\n",
              "      <td>7.751811</td>\n",
              "      <td>6.818331</td>\n",
              "      <td>...</td>\n",
              "      <td>-2.347892</td>\n",
              "      <td>-3.573529</td>\n",
              "      <td>-6.638988</td>\n",
              "      <td>-5.852189</td>\n",
              "      <td>8.638922</td>\n",
              "      <td>-3.788154</td>\n",
              "      <td>-4.986428</td>\n",
              "      <td>-6.144402</td>\n",
              "      <td>-2.663138</td>\n",
              "      <td>-0.227634</td>\n",
              "      <td>0.630000</td>\n",
              "      <td>-1.148500</td>\n",
              "      <td>-0.321750</td>\n",
              "      <td>-0.875137</td>\n",
              "      <td>-10.835594</td>\n",
              "      <td>1.012006</td>\n",
              "      <td>0.616927</td>\n",
              "      <td>-0.123525</td>\n",
              "      <td>0.815517</td>\n",
              "      <td>1.417692</td>\n",
              "      <td>1.730117</td>\n",
              "      <td>2.003574</td>\n",
              "      <td>2.723122</td>\n",
              "      <td>1.179165</td>\n",
              "      <td>2.852669</td>\n",
              "      <td>39.458965</td>\n",
              "      <td>5.736304</td>\n",
              "      <td>-20.788565</td>\n",
              "      <td>-5.314080</td>\n",
              "      <td>0.397473</td>\n",
              "      <td>1.973049</td>\n",
              "      <td>4.607931</td>\n",
              "      <td>7.300556</td>\n",
              "      <td>9.865562</td>\n",
              "      <td>3.085788</td>\n",
              "      <td>1.254327</td>\n",
              "      <td>1.854584</td>\n",
              "      <td>2.621457</td>\n",
              "      <td>-24.944792</td>\n",
              "      <td>1.078750</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>chr1</td>\n",
              "      <td>930336</td>\n",
              "      <td>843786</td>\n",
              "      <td>G</td>\n",
              "      <td>A</td>\n",
              "      <td>13.966989</td>\n",
              "      <td>-14.674191</td>\n",
              "      <td>-17.506096</td>\n",
              "      <td>10.756916</td>\n",
              "      <td>-1.225523</td>\n",
              "      <td>2.911127</td>\n",
              "      <td>0.152567</td>\n",
              "      <td>-3.519540</td>\n",
              "      <td>0.049325</td>\n",
              "      <td>-11.031021</td>\n",
              "      <td>1.868334</td>\n",
              "      <td>1.091527</td>\n",
              "      <td>-2.538539</td>\n",
              "      <td>0.955972</td>\n",
              "      <td>1.841750</td>\n",
              "      <td>-4.661942</td>\n",
              "      <td>-13.774675</td>\n",
              "      <td>-9.152153</td>\n",
              "      <td>-0.745051</td>\n",
              "      <td>6.185897</td>\n",
              "      <td>1.756672</td>\n",
              "      <td>-18.894367</td>\n",
              "      <td>5.834339</td>\n",
              "      <td>-8.249739</td>\n",
              "      <td>-11.984449</td>\n",
              "      <td>-11.462355</td>\n",
              "      <td>1.577443</td>\n",
              "      <td>-6.717025</td>\n",
              "      <td>0.324789</td>\n",
              "      <td>-4.330544</td>\n",
              "      <td>-2.505473</td>\n",
              "      <td>0.463757</td>\n",
              "      <td>7.450192</td>\n",
              "      <td>7.164897</td>\n",
              "      <td>4.730538</td>\n",
              "      <td>...</td>\n",
              "      <td>-19.131910</td>\n",
              "      <td>-4.917017</td>\n",
              "      <td>-8.311162</td>\n",
              "      <td>-5.551866</td>\n",
              "      <td>7.141852</td>\n",
              "      <td>-4.919424</td>\n",
              "      <td>-3.647799</td>\n",
              "      <td>-6.864504</td>\n",
              "      <td>-2.496408</td>\n",
              "      <td>-9.047115</td>\n",
              "      <td>-3.398409</td>\n",
              "      <td>-6.225352</td>\n",
              "      <td>-2.351780</td>\n",
              "      <td>-2.361993</td>\n",
              "      <td>-17.526522</td>\n",
              "      <td>-46.520599</td>\n",
              "      <td>-57.027424</td>\n",
              "      <td>-8.135074</td>\n",
              "      <td>-25.883907</td>\n",
              "      <td>-6.518816</td>\n",
              "      <td>-1.284766</td>\n",
              "      <td>-2.227928</td>\n",
              "      <td>-1.223655</td>\n",
              "      <td>-1.058415</td>\n",
              "      <td>3.015599</td>\n",
              "      <td>69.669312</td>\n",
              "      <td>-6.541300</td>\n",
              "      <td>-73.421532</td>\n",
              "      <td>-36.244965</td>\n",
              "      <td>-20.472229</td>\n",
              "      <td>-8.516649</td>\n",
              "      <td>8.461418</td>\n",
              "      <td>-3.483475</td>\n",
              "      <td>14.235535</td>\n",
              "      <td>6.582296</td>\n",
              "      <td>4.010679</td>\n",
              "      <td>1.865496</td>\n",
              "      <td>-6.477487</td>\n",
              "      <td>-42.006096</td>\n",
              "      <td>-0.880794</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 3378 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "  chrom     pos  ... CAGE:Smooth muscle c CAGE:parietal cortex\n",
              "0  chr1  930188  ...           -46.489288           -11.498157\n",
              "1  chr1  930203  ...            -4.407548             3.055573\n",
              "2  chr1  930248  ...             6.409196            -0.923778\n",
              "3  chr1  930275  ...           -24.944792             1.078750\n",
              "4  chr1  930336  ...           -42.006096            -0.880794\n",
              "\n",
              "[5 rows x 3378 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    }
  ]
}