{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "20e70aed-28af-41b4-87ad-389816ef8397",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ['CUDA_VISIBLE_DEVICES'] = '0'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "21c9df03-3840-4c82-b964-5ab94d23cd86",
   "metadata": {},
   "source": [
    "### Tutorial B5: Variant Effect\n",
    "\n",
    "A common use-case of sequence-based machine learning methods is predicting the effects of variants. By comparing the predictions of the model before and after incorporating a variant one can evaluate the predicted effect and, potentially, better understand the mechanism of its action. For example, let's pretend that we have a model that predicts the binding of some protein of interest. If this model predicts binding on a reference sequence but no longer predicts binding after including the variant, it would be reasonable to assume that (1) the variant could be involved in some downstream change in phenotype or (2) if the variant is already connected to some phenotype that the mechanism of its action might be through the blocking of the binding of this protein.  In contrast, if a variant does not change the predictions from the model, potentially it is not involved with the change in phenotype.\n",
    "\n",
    "There are many kinds of variants and some kinds are not trivial to calculate the effect for. The most basic kind are <i>substitutions</i>, where one or more characters is/are changed into another/other characters. A more complicated sort of variant is the <i>deletion</i>, where characters in one sequence are not present in the other sequence. Finally, there is the conceptual opposite of the deletion, the <i>insertion</i>, where characters are present in the other sequence but not the original sequence. Naturally, deletions and insertions are both relative, in the sense that what is viewed as a deletion in one set of sequences is just an insertion if you flip the reference. However, in genomics, we are usually using a reference genome and so deletions and insertions are distinguished according to the reference.\n",
    "\n",
    "Usually, calculating variant effect for substitutions is trivial because characters are simply swapped; however, calculating variant effect for insertions or deletions can be challenging because the two sequences are no longer the same length and most machine learning methods assume a fixed window length. Below, we will walk through what `tangermeme` does for each sort of variant."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d4d32fa-2035-413e-9c36-0f14206f3026",
   "metadata": {},
   "source": [
    "#### Substitutions\n",
    "\n",
    "The simplest form of variant is the substitution, where one or more characters are exchanged for other characters. These variants are the simplest to evaluate because one can simply run the original sequence through the model, make the substitutions, and run the new sequence through the model as well -- the original and modified sequences are both of the same length.\n",
    "\n",
    "We can use `tangermeme` to calculate these effects using the `substitution_effect` function. This function takes in a model, a set of original sequences (before inclusion of the variant), a COO-sparse matrix of substitutions, and optional additional arguments to pass into the provided function. Each line in the matrix of substitutions should be one substitution with the first column being the example, the second column being the position within the example, and the third column being the character that should appear at that position in that example. An important note is that this allows for multiple substitutions to be included in the same examples by having multiple lines with the same example index.\n",
    "\n",
    "For the purpose of demontrating the usage we can use our simple untrained dense model again."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "8759abb6-f4a4-42d7-962b-4a86b0a1907b",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from tangermeme.utils import random_one_hot\n",
    "\n",
    "class FlattenDense(torch.nn.Module):\n",
    "\tdef __init__(self, seq_len=10):\n",
    "\t\tsuper(FlattenDense, self).__init__()\n",
    "\t\tself.dense = torch.nn.Linear(seq_len*4, 3)\n",
    "\t\tself.seq_len = seq_len\n",
    "\n",
    "\tdef forward(self, X, alpha=0, beta=1):\n",
    "\t\tX = X.reshape(X.shape[0], self.seq_len*4)\n",
    "\t\treturn self.dense(X) * beta + alpha\n",
    "\n",
    "model = FlattenDense()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebb0c19c-02db-40d0-89b7-b96747480dd9",
   "metadata": {},
   "source": [
    "Let's start off by generating a random sequence."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "5083f7ba-2203-4aad-b5fb-4ec0bff3189c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tangermeme.utils import random_one_hot\n",
    "\n",
    "X = random_one_hot((1, 4, 10), random_state=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "25cc5a32-5e9b-4ab0-8a24-9d05440a8875",
   "metadata": {},
   "source": [
    "Next, let's create the substitutions. Let's say that we want to add in a `C` at position 4 (0-indexed) and a `G` at position 5."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "bde8576b-379e-4a3c-b885-74a407deb5f1",
   "metadata": {},
   "outputs": [],
   "source": [
    "substitutions = torch.tensor([\n",
    "    [0, 4, 1],   # Example index 0, position 4, C is index 1\n",
    "    [0, 5, 2]    # Example index 0, position 5, G is index 2\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d3a1173d-6899-4c8e-918d-81660af6c789",
   "metadata": {},
   "source": [
    "Now that we have our model, our original sequences, and our references, we can use the `substitution_effect` function to get the model predictions before and after making the substitutions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "90bb12a1-659d-4706-8315-82ecc970dc02",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 3]), torch.Size([1, 3]))"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tangermeme.variant_effect import substitution_effect\n",
    "\n",
    "y, y_var = substitution_effect(model, X, substitutions, device='cpu')\n",
    "y.shape, y_var.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1f791f2e-9896-4e96-9bdc-76bfd838ff2d",
   "metadata": {},
   "source": [
    "The tensors have the shape `(1, 3)` because there was only one example passed in and the effect is returned on the example level, not the individual variant level. This is an important point. If you have multiple variants whose marginal effect you would like to assess (i.e., you want to put each variant individually into the sequence and see what the predictions are) you will need to copy the example repeatedly and have a substitution matrix where each substitution goes into a different copy of the example.\n",
    "\n",
    "Further, `tangermeme` does not force the user to adopt a specific distance function. Rather, the functions returns the raw predictions with and without the substitutions and allows the user to define their own distance. For example, we could use a simple Euclidean distance function to calculate the difference between predictions before and after including the substitutions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "46965ff7-0a87-4681-a318-0454c68e6c33",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([0.2091])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.sqrt(torch.sum((y_var - y) ** 2, dim=-1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "47bbbc6b-f867-4f1d-87a7-10be8c850f45",
   "metadata": {},
   "source": [
    "More generally, these functions are meant to be the base operation that your downstream functions are built off of. In your library, you might write your own `variant_effect` function with a similar signature, call this function to get the predictions, calculate distance in the manner you choose, and return that -- either individually, or as part of the variant DataFrame that was passed in originally.\n",
    "\n",
    "##### GATA2\n",
    "\n",
    "As a more realistic example, let's look at what happens when we use a GATA2 BPNet model at a site that originally contains a GATA motif, but has to also consider a single substitution to that motif. Here, we will insert the GATA sequence into random background sequence just to quickly construct an example, and then in the substitution matrix swap one of the Ts to a C."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "475ba420-04cf-4d1a-aeb2-7f4171421e23",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2, torch.Size([1, 2, 1000]), torch.Size([1, 1]))"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tangermeme.ersatz import substitute\n",
    "\n",
    "bpnet = torch.load(\"gata2.bpnet.torch\")\n",
    "\n",
    "X3 = random_one_hot((1, 4, 2114), random_state=0).type(torch.float32)\n",
    "X3 = substitute(X3, \"CTTATCT\")\n",
    "\n",
    "substitutions3 = torch.tensor([\n",
    "    [0, 1058, 0]\n",
    "])\n",
    "\n",
    "y, y_var = substitution_effect(bpnet, X3, substitutions3, args=(torch.zeros(1, 2, 2114),), device='cpu')\n",
    "len(y), y[0].shape, y[1].shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d5642531-f38d-4cb0-bb05-e728934bb0fd",
   "metadata": {},
   "source": [
    "BPNet models return predicted probabilities for each basepair as well as the log counts for the region. Put another way, these models are multi-output, yet the variant effect functions are still able to handle these irregular shapes because the underlying predict function handles them well. For the purpose of this example we will just take a look a the predicted probabilities."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "2262196a-366d-427e-badf-46745663b028",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "import seaborn; seaborn.set_style('whitegrid')\n",
    "\n",
    "yb = torch.nn.functional.softmax(y[0][0], dim=-1) \n",
    "ya = torch.nn.functional.softmax(y_var[0][0], dim=-1)\n",
    "\n",
    "plt.figure(figsize=(12, 4))\n",
    "plt.subplot(121)\n",
    "plt.title(\"Before Variant Inclusion\")\n",
    "plt.plot(yb[0], color='r', linewidth=1, label=\"+ strand before\")\n",
    "plt.plot(yb[1], color='b', linewidth=1, label=\"- strand before\")\n",
    "plt.xlabel(\"Genomic Position\")\n",
    "plt.ylabel(\"Predicted Probability\")\n",
    "plt.legend(fontsize=8)\n",
    "\n",
    "plt.subplot(122)\n",
    "plt.title(\"After Variant Inclusion\")\n",
    "plt.plot(ya[0], color='0.3', linewidth=1, label=\"+ strand after\")\n",
    "plt.plot(ya[1], color='0.7', linewidth=1, label=\"- strand after\")\n",
    "plt.xlabel(\"Genomic Position\")\n",
    "plt.legend(fontsize=8)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e851a9b-2c1d-4f4b-9529-fac44e39ad02",
   "metadata": {},
   "source": [
    "It looks like that single T -> C variant causes the model to no longer predict a peak in the signal here. This is quite a dramatic change for just a single nucleotide and indicates that someone who had this variant might be worth investigating further."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e9c38469-cb0e-4ce5-be4a-319ffd1bb6b7",
   "metadata": {},
   "source": [
    "##### Arguments\n",
    "\n",
    "If you have additional arguments that need to be passed into the function you can do that through the use of the `args` argument. This argument takes in a tuple of tensors with the same first dimension as `X` and where the i-th entry in `X` gets processed with the i-th entry in each of the tensors in `args`. So, with our dense example, we can pass in tensors of alphas and betas as well."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9bdf1d07-b98f-44af-940d-60e774939ac5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([8, 3]), torch.Size([8, 3]))"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X2 = random_one_hot((8, 4, 10), random_state=0)\n",
    "alpha = torch.randn(8, 1)\n",
    "beta = torch.randn(8, 1)\n",
    "\n",
    "y, y_var = substitution_effect(model, X2, substitutions, args=(alpha, beta), device='cpu')\n",
    "y.shape, y_var.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "30340968-dd6f-47be-b560-3f0a5699bb8c",
   "metadata": {},
   "source": [
    "Here, we aren't actually making substitutions past the first example because we are using the same substitution matrix as before, but the point is to demonstrate how to pass additional arguments into the forward function of your model."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "04f7ba0c-b5ea-42e6-a1d8-31e957486f9c",
   "metadata": {},
   "source": [
    "##### Attributions\n",
    "\n",
    "By default, predictions are returned before and after the variants are included in the examples. However, like other functions in `tangermeme`, any function can be passed in to be applied before the variants are included. For example, if we wanted to get attributions before and after the substitutions we could do the following:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "78d42711-5513-4bef-aef6-c3c4de1b32d5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 4, 10]), torch.Size([1, 4, 10]))"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tangermeme.deep_lift_shap import deep_lift_shap\n",
    "\n",
    "y, y_var = substitution_effect(model, X, substitutions, func=deep_lift_shap, device='cpu')\n",
    "y.shape, y_var.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "05c91a51-f618-46b2-9905-ca4578a88c48",
   "metadata": {},
   "source": [
    "Note here that the shape is different because the model makes three predictions per example, yielding a shape of `(1, 3)`, but the attributions are one value for each input feature and so match the shape of `X`.\n",
    "\n",
    "Also like other functions, if we want to pass arguments into the `deep_lift_shap` function we can either simly pass them into `substitution_effect` or we can pass them as a dictionary into `additional_func_kwargs`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "6e27bde3-b7be-4cc8-a5d3-0a96a2bc0937",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 4, 10]), torch.Size([1, 4, 10]))"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y, y_var = substitution_effect(model, X, substitutions, func=deep_lift_shap, device='cpu', n_shuffles=2)\n",
    "y.shape, y_var.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "cef5556e-c861-4687-8e80-e8e23b6d67f5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 4, 10]), torch.Size([1, 4, 10]))"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y, y_var = substitution_effect(model, X, substitutions, func=deep_lift_shap, device='cpu', additional_func_kwargs={'n_shuffles': 2})\n",
    "y.shape, y_var.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f4e65bf8-3b96-4988-b28c-1cb89c77c30f",
   "metadata": {},
   "source": [
    "In case you missed it, this is to allow an option to completely ensure that arguments are passed solely into the function and do not get used by `substitution_effect` instead. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cdab2c5d-73a4-4fba-909c-e239699c0fee",
   "metadata": {},
   "source": [
    "#### Deletions\n",
    "\n",
    "Another form of variant is a deletion, where a character is removed from the sequence. Evaluating deletions is a little more challenging than substitutions because, with substitutions, every character remains the same except for the substitutions and -- more importantly -- the length of the sequence does not change. With a deletion, many characters will appear to change because they are being moved over by one or more positions. Afterward, the sequence with the deletions will be shorter than the original sequence, which poses challenges if the model expects sequences of the same length. Like with substitutions, multiple deletions can be specified for each sequence.\n",
    "\n",
    "These issues are overcome in `tangermeme` by, simply put, having the user load more sequence into each example and then slicing sequences to make them ultimately be the same length. More specifically, if the maximum number of deletions in one of the examples is 6 and the model expects a sequence of length 100 then the user must provide tensors with 106 positions for each example. This way, even after incorporating deletions in the sequence with the most deletions, all sequences will have at least the 100 positions the model expects.\n",
    "\n",
    "However, this does not completely solve the issue. First, the original sequences still all have 106 positions in them. Second, not all sequences will be incorporating 6 deletions and so will have different lengths. The solution to both of these problems involves trimming: trimming the original sequences down to 100 positions by removing 6 from an edge, and trimming a variable number of positions from each variant-incorporated sequence to get them down to 100 positions. Because one could trim from either the left or right side, the user specified `left=True` which trims from the left side or `left=False` which means trim from the right side."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "30d3c1da-32e1-4af8-bab8-5363c9b603d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = random_one_hot((1, 4, 12), random_state=0)   # Note that the length is longer than the expected 10\n",
    "\n",
    "deletions = torch.tensor([\n",
    "    [0, 1],   # Example index 0, delete the character at position 1\n",
    "    [0, 4]\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee67cc33-f950-4471-8fc9-0a7bd1ed8187",
   "metadata": {},
   "source": [
    "The function signature is identical as for substitutions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "84896e69-2696-4368-8a00-ed1ed7db805d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 3]), torch.Size([1, 3]))"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tangermeme.variant_effect import deletion_effect\n",
    "\n",
    "y, y_var = deletion_effect(model, X, deletions, device='cpu')\n",
    "y.shape, y_var.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f1f7d0f9-839a-48ad-a405-90f7062e736f",
   "metadata": {},
   "source": [
    "Like with substitutions, you get the predictions before and after the incorporation of the deletion. Similarly, you can also pass in your own custom function, additional tensors for the forward function, and other parameters."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6dfa13f3-cc22-4218-92cf-ad8a9aa11821",
   "metadata": {},
   "source": [
    "#### Insertions\n",
    "\n",
    "Finally, the conceptual opposite of the deletion is the insertion. Rather than removing a character from a sequence, you are adding a character, and so the same trimming strategy gets employed but sort of in reverse. Specifically, the user will pass in sequences of the expected length by the model and, after including characters according to the insertion matrix, characters from either the left or right will be trimmed. The original sequences are left untouched in this case."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "1c8b4b6c-d373-46c9-b8a5-0034db596a10",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 3]), torch.Size([1, 3]))"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from tangermeme.variant_effect import insertion_effect\n",
    "\n",
    "X = random_one_hot((1, 4, 10), random_state=0)   # Note that the length is longer than the expected 10\n",
    "\n",
    "insertions = torch.tensor([\n",
    "    [0, 1, 1],   # Example index 0, delete the character at position 1\n",
    "    [0, 4, 1]\n",
    "])\n",
    "\n",
    "y, y_var = insertion_effect(model, X, insertions, device='cpu')\n",
    "y.shape, y_var.shape"
   ]
  }
 ],
 "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}