{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "35ab8698-b3c8-4cf4-886e-a05859603978",
   "metadata": {},
   "source": [
    "### Custom Loss Functions\n",
    "\n",
    "Ledidi optimizes an objective function that is made up of two terms: the input loss, and the output loss. The input loss is, by default, an L1 loss between the original sequence and the proposed sequence and so effectively counts the number of edits that have been made. The output loss is, by default, an L2 loss between the model predictions given the edited sequence and the desired output from the model. Combined, this objective function tries to minimize the number of edits made in order to achieve a desired prediction from the model.\n",
    "\n",
    "However, the losses do not have to be these defaults. In general, using L1 as the input loss probably makes the most conceptual sense given the formulation of Ledidi, but there may be alternatives with better convergence properties. We anticipate that the most creativity will come from using custom loss functions for the output loss, where a great deal of flexibility can be used to control the design process or work with models with interesting outputs."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "460b0ef9-bdd0-4af0-b674-4e54038ec530",
   "metadata": {},
   "source": [
    "#### Handling models with structured outputs\n",
    "\n",
    "In most of our evaluations, we have either used models whose output takes the form of a count (e.g., Beluga, BPNet) or models whose output can be easily converted into something that acts like a count (e.g., Enformer, Borzoi). This is mostly because it is easy to work with models whose outputs are like this. But not all models can easily have their output converted into something that acts as a count, and even when they can, sometimes this eliminates useful information that is in the output.\n",
    "\n",
    "For example, BPNet models make predictions for both the total counts in the region but also the basepair resolution profiles of where these counts are predicted to map to. These profiles take the form of probability distributions where the output correspond to either the probability of a read mapping to each basepair, of something like an unnormalized logit that can be easily converted to such a profile. We have been discarding these profiles because they are harder to work with, but they also provide us with the potential to finely control *where* binding sites are inserted as opposed to just *how many*.\n",
    "\n",
    "Let's take a look. Here, we will quickly define a wrapper that pulls out the predicted profile *instead of* the predicted total counts, as we have been using before. Because the BPNet objects from bpnetlite predict logits by default, we will quickly convert those to probabilities in a numerically stable way using `log_softmax`. Note that the `ProfileWrapper` object that is built into bpnet-lite does not simply return the profile, but rather returns a conversion of the profile into a single number for use in attribution algorithms. tl;dr, do not simply use that one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "30c8cfcc-10e8-4a42-9128-383a88fed37c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "\n",
    "from bpnetlite.bpnet import ControlWrapper\n",
    "\n",
    "class ProfileWrapper(torch.nn.Module):\n",
    "    def __init__(self, model):\n",
    "        super(ProfileWrapper, self).__init__()\n",
    "        self.model = model\n",
    "    \n",
    "    def forward(self, X):\n",
    "        y = self.model(X)[0]\n",
    "        return torch.exp(torch.nn.functional.log_softmax(y, dim=-1))\n",
    "\n",
    "\n",
    "model = torch.load(\"../../../../models/bpnet/GATA2.torch\", weights_only=False)\n",
    "model = ProfileWrapper(ControlWrapper(model))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e4f8a2bf-696d-4bd5-90a5-7585f5e99582",
   "metadata": {},
   "source": [
    "Now that we have our model, let's generate a random sequence and apply our BPNet model to it to get a sense for what the profile predictions look like."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6c95c565-598e-47cf-8db1-f416d85622ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tangermeme.utils import random_one_hot\n",
    "from tangermeme.predict import predict\n",
    "\n",
    "X = random_one_hot((1, 4, 2114), random_state=0).float()\n",
    "y_hat = predict(model, X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d63ac95c-eb32-4f4c-8755-57718a05065b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "import seaborn; seaborn.set_style('whitegrid')\n",
    "\n",
    "plt.figure(figsize=(8, 3))\n",
    "plt.plot(y_hat[0, 0], c='0.3', label=\"fwd strand\")\n",
    "plt.plot(y_hat[0, 1], c='0.7', label=\"bwd strand\")\n",
    "\n",
    "plt.title(\"Original BNet Predictions\")\n",
    "plt.xlabel(\"Genomic Position\")\n",
    "plt.ylabel(\"Predicted Probability\")\n",
    "\n",
    "plt.xlim(0, 1000)\n",
    "plt.legend()\n",
    "\n",
    "seaborn.despine(bottom=True, left=True)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0dcf556f-9a2e-4aea-b7c6-67eca04251e7",
   "metadata": {},
   "source": [
    "There are a few things to make note of here. Of course, the first is that we are no longer dealing with a single number but rather one number per position in the 1kbp output. But it's not actually one number, it's two numbers, because the profile predictions from BPNet are stranded. An important property here is that the predictions will be offset from each other because BPNet was trained on fragment starts but the binding event happens in the middle of the fragment. This means that reads mapping to the + strand will map to the left of the actual binding event and reads mapping to the - strand will map to the right of the actual binding event and that the distance to the binding event is proportional to the fragment length, which is expertiment-specific. Finally, each strand has been normalized separately and the predicted probabilities for each strand sum to 1.\n",
    "\n",
    "These factors matter when creating the desired output from the model. We can begin by specifying a tensor in a conceptually similar manner to when we worked with multiple outputs. After all, one can view predictions for each basepair as basically the same thing as predictions from multiple models. But when we set the desired output, we need to keep in mind an appropriate magnitude for these predictions (a value of 2 at a certain basepair is infeasible if the sum across the region has to be equal to 1) and also that we expect the peaks to be offset across the two strands.\n",
    "\n",
    "Fortunately, these are both somewhat easy to handle. We can just specify where we want a peak to be and calculate spans from there, making sure to divide by the total at the end so the desired output sums to 1 across each strand."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a6846049-8972-4853-b222-609e43a05eb9",
   "metadata": {},
   "outputs": [],
   "source": [
    "mid = 250\n",
    "\n",
    "y_bar = torch.zeros(1, 2, 1000).float()\n",
    "y_bar[:, 0, mid-150:mid+50] = 1\n",
    "y_bar[:, 1, mid-50 :mid+150] = 1\n",
    "y_bar /= y_bar.sum(dim=-1, keepdims=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "739a0be2-3396-4bab-b5c5-00782195cbb9",
   "metadata": {},
   "source": [
    "Given this desired output we can just plop it into Ledidi! As a technical note, given how small the predicted and desired values are, we would need to drop lambda by quite a bit to get the input and output losses to be on the same scale. For the purpose of demonstration we will just drop lambda."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "11fa2dcc-a8a6-4090-a0a6-e9faf05c68b7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=4.414e-06\ttotal_loss=4.414e-06\ttime=0.0\n",
      "iter=100\tinput_loss=38.38\toutput_loss=1.457e-06\ttotal_loss=1.457e-06\ttime=2.501\n",
      "iter=200\tinput_loss=39.56\toutput_loss=9.372e-07\ttotal_loss=9.372e-07\ttime=1.196\n",
      "iter=300\tinput_loss=38.75\toutput_loss=8.945e-07\ttotal_loss=8.945e-07\ttime=1.195\n",
      "iter=400\tinput_loss=40.12\toutput_loss=8.436e-07\ttotal_loss=8.436e-07\ttime=1.195\n",
      "iter=F\tinput_loss=42.19\toutput_loss=8.31e-07\ttotal_loss=8.31e-07\ttime=7.136\n"
     ]
    }
   ],
   "source": [
    "from ledidi import ledidi\n",
    "\n",
    "X_hat = ledidi(model, X, y_bar, l=0.0, verbose=True)\n",
    "y_hat = predict(model, X_hat)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71fd491a-68f3-4805-9405-b7be05ef561c",
   "metadata": {},
   "source": [
    "We did not need to do anything particularly special here to work with the profiles from Ledidi. Let's take a look at the design, noting where we wanted the peaks to happen."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "067ff632-c102-47c4-82ff-6c0beda744be",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "import seaborn; seaborn.set_style('whitegrid')\n",
    "\n",
    "plt.figure(figsize=(8, 3))\n",
    "plt.plot(y_hat[0, 0], c='0.3', label=\"fwd strand\")\n",
    "plt.plot(y_hat[0, 1], c='0.7', label=\"bwd strand\")\n",
    "\n",
    "plt.plot([100, 300], [0.0090, 0.0090], c='0.3', linestyle='--', linewidth=2)\n",
    "plt.plot([200, 400], [0.0087, 0.0087], c='0.7', linestyle='--', linewidth=2)\n",
    "\n",
    "plt.title(\"Ledidi Design using MSE\")\n",
    "plt.xlabel(\"Genomic Position\")\n",
    "plt.ylabel(\"Predicted Probability\")\n",
    "\n",
    "plt.xlim(0, 1000)\n",
    "plt.legend()\n",
    "\n",
    "seaborn.despine(bottom=True, left=True)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4c6e5b1f-1f41-4854-a066-ca2fee685f5f",
   "metadata": {},
   "source": [
    "Great! Looks like we have peak predictions at the desired locations.\n",
    "\n",
    "But is MSE the best loss to optimize here? BPNet intentionally chooses to use a different loss when optimizing the profile predictions because MSE has been shown to have weird artifacts when working with probabilities, like this. For example, note how much density of the predictions falls out of the desired range.\n",
    "\n",
    "Maybe we can do better if we use a loss function that is explicitly meant to operate on probability distributions. KL-divergence has been shown to work well in this setting. To use a custom loss function, all we need to do is pass in a function to the `output_loss` parameter with the signature `f(y_hat, y_bar)` where `y_bar` is the desired output and `y_hat` is the predicted output given the current edited sequence. As a technical note, we are wrapping it a little bit because the implementation we are using reports the KL divergence for each example but we need the average divergence across all examples."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "01071969-18f7-45e9-afc1-9502d6ff3750",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=1.755\ttotal_loss=1.755\ttime=0.0\n",
      "iter=100\tinput_loss=751.0\toutput_loss=0.1578\ttotal_loss=0.1578\ttime=1.26\n",
      "iter=200\tinput_loss=702.8\toutput_loss=0.1326\ttotal_loss=0.1326\ttime=1.282\n",
      "iter=300\tinput_loss=760.6\toutput_loss=0.1126\ttotal_loss=0.1126\ttime=1.258\n",
      "iter=400\tinput_loss=762.8\toutput_loss=0.09138\ttotal_loss=0.09138\ttime=1.258\n",
      "iter=500\tinput_loss=776.6\toutput_loss=0.08533\ttotal_loss=0.08533\ttime=1.262\n",
      "iter=600\tinput_loss=793.1\toutput_loss=0.07705\ttotal_loss=0.07705\ttime=1.254\n",
      "iter=700\tinput_loss=800.4\toutput_loss=0.07333\ttotal_loss=0.07333\ttime=1.256\n",
      "iter=800\tinput_loss=823.8\toutput_loss=0.07241\ttotal_loss=0.07241\ttime=1.255\n",
      "iter=F\tinput_loss=807.0\toutput_loss=0.06802\ttotal_loss=0.06802\ttime=10.29\n"
     ]
    }
   ],
   "source": [
    "from bpnetlite.performance import _kl_divergence\n",
    "\n",
    "kld = lambda x, y: _kl_divergence(y, x).mean()\n",
    "\n",
    "X_hat = ledidi(model, X, y_bar, l=0.0, output_loss=kld, verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "bc5db550-f78d-40fe-99bf-a9e6dfa9ddb3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_hat = predict(model, X_hat)\n",
    "\n",
    "plt.figure(figsize=(6, 3))\n",
    "plt.plot(y_hat[0, 0], c='0.3', label=\"fwd strand\")\n",
    "plt.plot(y_hat[0, 1], c='0.7', label=\"bwd strand\")\n",
    "\n",
    "plt.title(\"Ledidi Design using KL divergence\")\n",
    "plt.xlabel(\"Genomic Position\")\n",
    "plt.ylabel(\"Predicted Probability\")\n",
    "\n",
    "plt.plot([100, 300], [0.0090, 0.0090], c='0.3', linestyle='--', linewidth=2)\n",
    "plt.plot([200, 400], [0.0087, 0.0087], c='0.7', linestyle='--', linewidth=2)\n",
    "\n",
    "plt.xlim(0, 1000)\n",
    "plt.legend()\n",
    "\n",
    "seaborn.despine(bottom=True, left=True)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bfcc3de9-5f7c-43f9-87ba-aa166191289f",
   "metadata": {},
   "source": [
    "That looks remarkably sharper. Not only are both peaks at basically the same height, but they sharply decrease in predictions at the desired edges. It is likely, especially given the number of edits necessary to make these profiles, that the designed sequence will look sort of weird. But this is just a demonstration of (1) how to use a custom loss fnction and (2) the differences in output one might expect when using a loss function more tailored for the model output.\n",
    "\n",
    "Let's try to take this one step further. Can we design edits that create two peaks where one peak is much stronger than the other?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e48780da-5315-431b-8816-f6f8f125db18",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=1.398\ttotal_loss=1.398\ttime=0.0\n",
      "iter=100\tinput_loss=718.0\toutput_loss=0.1765\ttotal_loss=0.1765\ttime=1.259\n",
      "iter=200\tinput_loss=722.9\toutput_loss=0.1369\ttotal_loss=0.1369\ttime=1.259\n",
      "iter=300\tinput_loss=748.1\toutput_loss=0.1254\ttotal_loss=0.1254\ttime=1.258\n",
      "iter=400\tinput_loss=778.4\toutput_loss=0.1056\ttotal_loss=0.1056\ttime=1.256\n",
      "iter=500\tinput_loss=757.3\toutput_loss=0.1188\ttotal_loss=0.1188\ttime=1.257\n",
      "iter=F\tinput_loss=763.6\toutput_loss=0.09379\ttotal_loss=0.09379\ttime=7.296\n"
     ]
    }
   ],
   "source": [
    "y_bar = torch.zeros(1, 2, 1000).float()\n",
    "y_bar[:, 0, 100:300] = 1\n",
    "y_bar[:, 1, 200:400] = 1\n",
    "y_bar[:, 0, 600:800] = 3\n",
    "y_bar[:, 1, 700:900] = 3\n",
    "y_bar /= y_bar.sum(dim=-1, keepdims=True)\n",
    "\n",
    "X_hat = ledidi(model, X, y_bar, l=0.0, output_loss=kld, verbose=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "44d22409-8fe2-4611-a1a9-1e3b02ff89a4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_hat = predict(model, X_hat)\n",
    "\n",
    "plt.figure(figsize=(8, 3))\n",
    "plt.plot(y_hat[0, 0], c='0.3', label=\"fwd strand\")\n",
    "plt.plot(y_hat[0, 1], c='0.7', label=\"bwd strand\")\n",
    "\n",
    "plt.title(\"Ledidi Design using KL divergence\")\n",
    "plt.xlabel(\"Genomic Position\")\n",
    "plt.ylabel(\"Predicted Probability\")\n",
    "\n",
    "plt.xlim(0, 1000)\n",
    "plt.legend()\n",
    "\n",
    "seaborn.despine(bottom=True, left=True)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34ec19b3-80a4-4bb8-be26-c44f650b5903",
   "metadata": {},
   "source": [
    "Looks like the answer is yes! Just because we have changed the output function to be KL divergence instead of MSE does not mean that we lose any flexibility to design edits with whatever properties we would like."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "febe0a5d-3022-4373-9127-29507b96c023",
   "metadata": {},
   "source": [
    "#### Unconstrained Rewards\n",
    "\n",
    "A downside of using MSE in the multi-output setting is that we have to have a precise target value that we want to match for each of the outputs. Sometimes, this is the case and so using MSE is not a problem. However, other times, we may not have or even be able to reasonably guess what the output should be from each of the models. In these cases, we usually have a general sense that we would like designs that generally increase or decrease predictions from the models but do not know what the exact values should be. \n",
    "\n",
    "As a concrete example, consider the design of accessible sites whose accessibility is driven by a specific TF. In this setting, we likely have a target accessibility value (e.g., choosing a target that makes the region in the top percentile of accessible sites). But we may not know what the corresponding predictions should be for the TF model. This is because the dynamic range for each TF model will depend on the read depth and quality of the data used to train it and also because the effect on accessibility of the binding of each TF differs based on the structure and activity of the protein. For example, a pioneer factor will open chromatin significantly more than a \"settler\" factor. Trying to estimate the connection between TF binding and accessibility and then converting the desired accessibility value into a TF binding value is tricky: set the value too low and other TFs will be used to drive accessibility, set the value too high and suddenly you have an infeasible objective where either accessibility will be too high or TF binding will be too low.\n",
    "\n",
    "A solution to this problem is to have an output loss function where some of the outputs have target values and other outputs are just rewarded for being higher or lower without constraint. In the previous example, we set a target accessibility value and then just reward higher predictions from the TF binding model without requiring that they meet a certain value. One way of viewing this is that there are many ways to design an accessible site and this approach pushes the designer towards those that utilize the binding of a specific TF.\n",
    "\n",
    "To see this in action, let's load up a ChromBPNet model and a BPNet model that makes predictions for the binding of MAX."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "71f544e2-cd0a-4585-851f-cb553c461533",
   "metadata": {},
   "outputs": [],
   "source": [
    "from bpnetlite import BPNet\n",
    "from bpnetlite.bpnet import CountWrapper\n",
    "\n",
    "X = random_one_hot((1, 4, 2114), random_state=0).float()\n",
    "\n",
    "chrombpnet = BPNet.from_chrombpnet(\"../../../../models/chrombpnet/fold_0/model.chrombpnet_nobias.fold_0.ENCSR868FGK.h5\")\n",
    "chrombpnet = CountWrapper(chrombpnet).cuda()\n",
    "\n",
    "bpnet_max = torch.load(\"../../../../models/bpnet/MAX.torch\", weights_only=False)\n",
    "bpnet_max = CountWrapper(ControlWrapper(bpnet_max))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aa7786ad-a512-453b-ba56-f8fa78f2c3aa",
   "metadata": {},
   "source": [
    "We can combine the two models and use them with Ledidi to design edits, just as we did in the multi-output tutorial."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "e9e756d3-75e4-4b45-a950-cfd130b1ffe6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=26.07\ttotal_loss=26.07\ttime=0.0\n",
      "iter=100\tinput_loss=108.4\toutput_loss=1.235\ttotal_loss=12.08\ttime=46.12\n",
      "iter=200\tinput_loss=121.2\toutput_loss=0.6067\ttotal_loss=12.73\ttime=24.36\n",
      "iter=300\tinput_loss=114.0\toutput_loss=0.773\ttotal_loss=12.17\ttime=24.37\n",
      "iter=F\tinput_loss=99.06\toutput_loss=0.8359\ttotal_loss=10.74\ttime=115.3\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([12.4062,  5.8811])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from ledidi.wrappers import DesignWrapper\n",
    "\n",
    "designer = DesignWrapper([chrombpnet, bpnet_max])\n",
    "\n",
    "y_bar = torch.tensor([[13.0, 7.0]])\n",
    "\n",
    "X_bar = ledidi(designer, X, y_bar, verbose=True)\n",
    "y_hat = predict(designer, X_bar)\n",
    "y_hat.mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4cde14de-5505-4e99-80c6-702530ae257b",
   "metadata": {},
   "source": [
    "Looks like it is able to generate edits that gets close to the two targets.\n",
    "\n",
    "Now, let's consider the alternate loss function. In this loss, we want to meet a desired output but we will also reward increases in MAX prediction. There are a few ones that one could implement this but probably the simplest is a function that takes in a `y_hat` tensor with two elements (accessibility and MAX predictions) and a `y_bar` tensor with one element (just the desired accessibility output), calculate a MSE loss on the accessibility component, and simply reward higher predictions from the MAX model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "91b82be5-509f-4a5d-a2c1-664821f23ab2",
   "metadata": {},
   "outputs": [],
   "source": [
    "def mse_max_loss(y_hat, y_bar):\n",
    "    return torch.nn.MSELoss()(y_hat[: :1], y_bar) - y_hat[:, 1].mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69c3a13e-1804-4416-bd2f-d220693118d1",
   "metadata": {},
   "source": [
    "A convenient aspect of framing the loss like this is that we do not have to make up some values for the \"desired output\" of MAX which we would then subsequently ignore. Rather, we are only passing in desired values for components that we actually want to control. Note that we are subtracting the MAX predictions (the second column in `y_hat`) because Ledidi minimizes the output loss and we want to maximize these predictions. Also note that we do not have to use the built-in MSE loss for the first component. As long as the function is differentiable we can do whatever we want. We are just using it here for convneience.\n",
    "\n",
    "After specifying the output loss function, we can just pass that in like we did before."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "3e6bf207-b671-4bf0-9c38-696ed0136cae",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=76.21\ttotal_loss=76.21\ttime=0.0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/users/jacob.schreiber/anaconda3/lib/python3.12/site-packages/torch/nn/modules/loss.py:608: UserWarning: Using a target size (torch.Size([1, 1])) that is different to the input size (torch.Size([1, 2])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.\n",
      "  return F.mse_loss(input, target, reduction=self.reduction)\n",
      "/users/jacob.schreiber/anaconda3/lib/python3.12/site-packages/torch/nn/modules/loss.py:608: UserWarning: Using a target size (torch.Size([16, 1])) that is different to the input size (torch.Size([16, 2])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.\n",
      "  return F.mse_loss(input, target, reduction=self.reduction)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=100\tinput_loss=326.2\toutput_loss=1.18\ttotal_loss=33.81\ttime=24.38\n",
      "iter=200\tinput_loss=360.9\toutput_loss=-4.321\ttotal_loss=31.77\ttime=24.39\n",
      "iter=300\tinput_loss=354.4\toutput_loss=-4.972\ttotal_loss=30.47\ttime=24.37\n",
      "iter=400\tinput_loss=347.4\toutput_loss=-5.126\ttotal_loss=29.62\ttime=24.38\n",
      "iter=500\tinput_loss=344.4\toutput_loss=-6.973\ttotal_loss=27.47\ttime=24.38\n",
      "iter=600\tinput_loss=335.4\toutput_loss=-7.331\ttotal_loss=26.21\ttime=24.38\n",
      "iter=F\tinput_loss=329.9\toutput_loss=-7.888\ttotal_loss=25.11\ttime=164.6\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([12.3439, 10.7223])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_bar = ledidi(designer, X, y_bar[:, :1], output_loss=mse_max_loss, verbose=True)\n",
    "\n",
    "y_hat = predict(designer, X_bar)\n",
    "y_hat.mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af9dec46-a07e-4785-bd49-81d4fe25edc5",
   "metadata": {},
   "source": [
    "By using this loss function we are able to get much higher predictions of MAX while still getting the same accessibility predictions. Because the binding of many TFs (including MAX) subsequently cause an increase in accessibility, one can infer that having much higher predictions for MAX mean that other TFs must be binding less to achieve the same accessibility. Although we will not go deeper into it here, one can augment this approach by additionally using models for any of TF that might be relevant and rewarding low predictions from them. Together, that would yield designs that explicitly want higher MAX predictions and lower predictions of other TFs."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "df299dfe-5e77-4cd1-9c19-68f493608e36",
   "metadata": {},
   "source": [
    "#### Cell Type-Specific Design and Min Gap\n",
    "\n",
    "A common goal when designing elements is trying to create cell type-specific activity. This can be achieved in the default setting by setting high predicted values in the cell type(s) that one wants the element to be active in and low values in the cell type(s) that the element should not be active in. However, there can be two issues with simply using MSE to design specific elements like this. First, because MSE has a smoothing property (like we saw at the beginning of the notebok) is may not yield elements that are truly as specific as one might like. Second, an increasing amount of recent work has shown that the most cell type-specific elements are those that only exhibit weak signal and that making these regions stronger might risk making them less specific. Put another way: the target value for a cell type-specific element might be on a different range than the normal dynamic range of the model.\n",
    "\n",
    "A potential solution to this wa proposed by Gosai <i>et al.</i> in what they called the \"min gap.\" Essentially, the loss function involves maximizing the gap between (1) the minimum on-target value and (2) the maximum off-target value. A strength of this loss function is that it gets around the smoothing property of MSE. It does not matter if the elements are off in 9 out of 10 off-target cell lines. Each component of the loss is only as good as the <i>worse</i> element in that group.\n",
    "\n",
    "Let's demonstrate this using the Malinois model from the same work."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "4df9c159-8c03-4bdc-9291-cfe9ac02540e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from boda.model import BassetBranched\n",
    "\n",
    "checkpoint = torch.load(\"../../../../models/malinois/torch_checkpoint.pt\", weights_only=False)\n",
    "malinois = BassetBranched(**vars(checkpoint['model_hparams']))\n",
    "malinois.load_state_dict(checkpoint['model_state_dict'])\n",
    "\n",
    "X = random_one_hot((1, 4, 600), random_state=0).float()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "18436f0a-7f55-4e25-a44d-162edc364289",
   "metadata": {},
   "source": [
    "We can load the MinGap loss from `ledidi.losses`. Because there are two groups -- the on-target and the off-target outputs -- we have to pass in a binary vector indicating which elements are part of the on-target group. To support any composition of classes, `MinGap` is a `torch.nn.Module` instance that you pass a binary vector into and whose forward function is defned as the loss function. Here, let's say that we want to design edits that turn the sequence into something specific to the second output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "2067b4d6-78f6-4e6a-9b33-8d98609f590f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from ledidi.losses import MinGap\n",
    "\n",
    "in_mask = torch.tensor([False, True, False])\n",
    "output_loss = MinGap(in_mask)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b9c7359c-44b9-4d7e-92dc-f016b7bad34a",
   "metadata": {},
   "source": [
    "An important aspect of the min gap loss is that there are no target values. This has two practical ramifications. First, there is nothing saying that the predictions have to be high in the on-target cell lines and low in the off-target ones. Because the loss only cares about maximizing the loss, both values could in theory be negative so long as the gap between them is high. Second, Ledidi's implementation expects something for `y_bar` even if that gets ignored by the loss. Here, we have to just make up some values but <i>they do not get used by min gap at all</i>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "b87a9fe7-997c-47a8-87fd-28724f38f8be",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=-0.1533\ttotal_loss=-0.1533\ttime=0.0\n",
      "iter=100\tinput_loss=92.38\toutput_loss=-21.22\ttotal_loss=-11.98\ttime=0.8177\n",
      "iter=200\tinput_loss=117.4\toutput_loss=-37.31\ttotal_loss=-25.57\ttime=0.6526\n",
      "iter=300\tinput_loss=123.6\toutput_loss=-38.47\ttotal_loss=-26.11\ttime=0.6503\n",
      "iter=400\tinput_loss=124.4\toutput_loss=-40.66\ttotal_loss=-28.23\ttime=0.6506\n",
      "iter=500\tinput_loss=120.2\toutput_loss=-41.29\ttotal_loss=-29.27\ttime=0.6505\n",
      "iter=600\tinput_loss=122.0\toutput_loss=-41.84\ttotal_loss=-29.64\ttime=0.6471\n",
      "iter=700\tinput_loss=124.6\toutput_loss=-42.18\ttotal_loss=-29.72\ttime=0.6456\n",
      "iter=800\tinput_loss=128.4\toutput_loss=-43.39\ttotal_loss=-30.55\ttime=0.6451\n",
      "iter=900\tinput_loss=126.2\toutput_loss=-42.93\ttotal_loss=-30.31\ttime=0.6477\n",
      "iter=F\tinput_loss=125.4\toutput_loss=-43.88\ttotal_loss=-31.33\ttime=6.318\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([ 2.5212, 46.3993,  2.4737])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_bar = torch.tensor([[0, 0, 0]])\n",
    "X_bar = ledidi(malinois, X, y_bar, output_loss=output_loss, verbose=True)\n",
    "\n",
    "y_hat = predict(malinois, X_bar)\n",
    "y_hat.mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c22f5510-6aad-4aef-b5c5-290032317c6d",
   "metadata": {},
   "source": [
    "Looks like the predictions from the second output are significantly higher than the predictions from the other two outputs.\n",
    "\n",
    "To show how general-purpose this loss is, let's say we want to make elements that are active in the first two but not the last output."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "45de60f3-2f35-4281-826b-d46e39880653",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=-0.2093\ttotal_loss=-0.2093\ttime=0.0\n",
      "iter=100\tinput_loss=66.06\toutput_loss=-12.86\ttotal_loss=-6.255\ttime=0.6808\n",
      "iter=200\tinput_loss=77.12\toutput_loss=-14.54\ttotal_loss=-6.827\ttime=0.7468\n",
      "iter=300\tinput_loss=85.38\toutput_loss=-15.81\ttotal_loss=-7.273\ttime=0.6467\n",
      "iter=F\tinput_loss=82.06\toutput_loss=-15.91\ttotal_loss=-7.704\ttime=2.674\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([16.6600, 16.7327,  0.7471])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_bar = torch.tensor([[0, 0, 0]])\n",
    "X_bar = ledidi(malinois, X, y_bar, output_loss=MinGap(torch.tensor([True, True, False])), verbose=True)\n",
    "\n",
    "y_hat = predict(malinois, X_bar)\n",
    "y_hat.mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fbf7ce8c-4d54-40a9-baf5-32b2d512c82e",
   "metadata": {},
   "source": [
    "#### Assay-specific Design and Min Gap\n",
    "\n",
    "As a side-note about the min gap loss, you do not have to use it only for cell type-specific design. More broadly, it can be used for <i>output-specific design</i> regardless of what those outputs are. If we combine two of our previous examples, let's say that we want to design elements that are TF specific. We can apply the min gap across TFs by designating them as either being on-target or off-target and design elements where the predictions should only be high for the on-target outputs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "3e5c91f0-9849-478a-a476-68cb263a837f",
   "metadata": {},
   "outputs": [],
   "source": [
    "bpnet_ctcf = torch.load(\"../../../../models/bpnet/CTCF.torch\", weights_only=False)\n",
    "bpnet_ctcf = CountWrapper(ControlWrapper(bpnet_ctcf))\n",
    "\n",
    "bpnet_e2f6 = torch.load(\"../../../../models/bpnet/E2F6.torch\", weights_only=False)\n",
    "bpnet_e2f6 = CountWrapper(ControlWrapper(bpnet_e2f6))\n",
    "\n",
    "bpnet_junb = torch.load(\"../../../../models/bpnet/JUNB.torch\", weights_only=False)\n",
    "bpnet_junb = CountWrapper(ControlWrapper(bpnet_junb))\n",
    "\n",
    "designer = DesignWrapper([bpnet_max, bpnet_ctcf, bpnet_e2f6, bpnet_junb])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "816d72bb-eaf2-4aff-9ceb-f357289951c5",
   "metadata": {},
   "source": [
    "Just as before we can specify a min gap loss according to these four outputs. If we want to try designing something that is CTCF-specific all we need to do is set it to be the in-group and everything else to be the out-group."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "8d6742e2-167a-47a5-bfd5-859930575b5d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=4.06\ttotal_loss=4.06\ttime=0.0\n",
      "iter=100\tinput_loss=744.4\toutput_loss=-4.478\ttotal_loss=-3.734\ttime=5.709\n",
      "iter=200\tinput_loss=868.8\toutput_loss=-6.906\ttotal_loss=-6.038\ttime=5.703\n",
      "iter=300\tinput_loss=869.7\toutput_loss=-7.69\ttotal_loss=-6.82\ttime=5.703\n",
      "iter=400\tinput_loss=878.2\toutput_loss=-8.609\ttotal_loss=-7.731\ttime=5.704\n",
      "iter=500\tinput_loss=911.9\toutput_loss=-9.325\ttotal_loss=-8.414\ttime=5.717\n",
      "iter=600\tinput_loss=932.3\toutput_loss=-9.352\ttotal_loss=-8.42\ttime=5.703\n",
      "iter=700\tinput_loss=970.6\toutput_loss=-9.956\ttotal_loss=-8.985\ttime=5.702\n",
      "iter=800\tinput_loss=996.0\toutput_loss=-9.996\ttotal_loss=-9.0\ttime=5.706\n",
      "iter=F\tinput_loss=1.005e+03\toutput_loss=-10.18\ttotal_loss=-9.174\ttime=50.97\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([ 1.3769, 14.2045,  3.2331,  4.0257])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = random_one_hot((1, 4, 2114), random_state=0).float()\n",
    "\n",
    "y_bar = torch.tensor([[0, 0, 0]])\n",
    "X_bar = ledidi(designer, X, y_bar, l=0.001, output_loss=MinGap(torch.tensor([False, True, False, False])), verbose=True)\n",
    "\n",
    "y_hat = predict(designer, X_bar)\n",
    "y_hat.mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa8568cd-9a93-4ca5-bd76-7c4994fe8e23",
   "metadata": {},
   "source": [
    "Looks like we are getting very strong predictions from the CTCF model and lower predictions from everything else.\n",
    "\n",
    "We can also specify groups of proteins that we would like to have high values just by setting several values to True."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "480b2949-e4e5-448e-900c-8a3a89bdc288",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=-0.9792\ttotal_loss=-0.9792\ttime=0.0\n",
      "iter=100\tinput_loss=422.1\toutput_loss=-5.737\ttotal_loss=-5.315\ttime=5.718\n",
      "iter=200\tinput_loss=548.4\toutput_loss=-6.782\ttotal_loss=-6.233\ttime=5.709\n",
      "iter=300\tinput_loss=598.2\toutput_loss=-7.302\ttotal_loss=-6.704\ttime=5.706\n",
      "iter=400\tinput_loss=625.4\toutput_loss=-7.944\ttotal_loss=-7.318\ttime=5.711\n",
      "iter=500\tinput_loss=655.7\toutput_loss=-8.127\ttotal_loss=-7.472\ttime=5.705\n",
      "iter=600\tinput_loss=670.2\toutput_loss=-8.68\ttotal_loss=-8.009\ttime=5.706\n",
      "iter=700\tinput_loss=663.0\toutput_loss=-8.955\ttotal_loss=-8.292\ttime=5.711\n",
      "iter=800\tinput_loss=660.1\toutput_loss=-9.117\ttotal_loss=-8.457\ttime=5.708\n",
      "iter=900\tinput_loss=661.3\toutput_loss=-9.458\ttotal_loss=-8.796\ttime=5.707\n",
      "iter=1000\tinput_loss=679.2\toutput_loss=-9.66\ttotal_loss=-8.981\ttime=5.707\n",
      "iter=F\tinput_loss=679.2\toutput_loss=-9.66\ttotal_loss=-8.981\ttime=57.09\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([-0.9855, -0.9598,  8.7180,  8.7110])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_bar = torch.tensor([[0, 0, 0]])\n",
    "X_bar = ledidi(designer, X, y_bar, l=0.001, output_loss=MinGap(torch.tensor([False, False, True, True])), verbose=True)\n",
    "\n",
    "y_hat = predict(designer, X_bar)\n",
    "y_hat.mean(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "169b2cee-28b2-4d3c-ad4b-c9efdcdda1ca",
   "metadata": {},
   "source": [
    "A note when using min gap in the multi-model setting is making sure that the dynamic range of predicted values between the models is comparable. This is important because it will be challenging (if not impossible) to get an off-target model to predict a smaller value than an on-target model when the range of the off-target model is between 10 and 20 and the range of the on-target model is between 0 and 4. Because experimental artfiacts such as read depth and quality will influence the range of the predictions, one may wish to carefully inspect the models and potentially even z-score normalize the predictions to circumvent these potential issues."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4bd71f8d-3f9b-45c6-939d-417001d3a34e",
   "metadata": {},
   "source": [
    "#### Custom Loss for Less Specificity\n",
    "\n",
    "A point that we have repeated about using MSE is that you have to set specific target values and that figuring out what these values should be can be challenging. One way to reduce the impact of this decision is to use a loss function that is more resistant to the precise target value and focuses more on getting in the right ballpark. As an example, we can define a custom \"ballpark loss\" that says a perfect match is anything within a radius of the target value. By using this loss, if you specify a desired value of 7.5, a predicted value of 7.2 is just as good as 7.5 -- you just want the predictions to be in the ballpark."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "ff8ff6e3-e286-4924-beec-febc78b1b0f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "def ballpark_loss(y_hat, y_bar, z=1):    \n",
    "    return torch.maximum(torch.nn.MSELoss(reduction='none')(y_hat, y_bar) - z, torch.zeros_like(y_hat))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16d0d806-6d13-43b3-8158-93c63e23968c",
   "metadata": {},
   "source": [
    "We can take a look at the loss directly to see what the different curves look like."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "10952449-f8a4-482a-ba81-500fbb1cdf23",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/users/jacob.schreiber/anaconda3/lib/python3.12/site-packages/torch/nn/modules/loss.py:608: UserWarning: Using a target size (torch.Size([1])) that is different to the input size (torch.Size([101])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.\n",
      "  return F.mse_loss(input, target, reduction=self.reduction)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y = torch.tensor([5.0])\n",
    "x = torch.arange(0, 10.1, 0.1)\n",
    "\n",
    "loss1 = torch.nn.MSELoss(reduction='none')(x, y)\n",
    "loss2 = ballpark_loss(x, y, 0.5).numpy(force=True)\n",
    "loss3 = ballpark_loss(x, y, 1).numpy(force=True)\n",
    "loss4 = ballpark_loss(x, y, 2).numpy(force=True)\n",
    "\n",
    "plt.title(\"Loss Functions (Target=5.0)\")\n",
    "plt.plot(x, loss1, label=\"MSE Loss\")\n",
    "plt.plot(x, loss2, label=\"Ballpark Loss (0.5)\")\n",
    "plt.plot(x, loss3, label=\"Ballpark Loss (1.0)\")\n",
    "plt.plot(x, loss4, label=\"Ballpark Loss (2.0)\")\n",
    "plt.xlabel(\"Prediction\")\n",
    "plt.ylabel(\"Loss\")\n",
    "\n",
    "plt.legend()\n",
    "seaborn.despine(bottom=True, left=True)\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e427dac0-35f1-498d-a7e7-b153429c2595",
   "metadata": {},
   "source": [
    "We can just plug the loss directly into Ledidi with the small modification that we need to take the mean to get a single number."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "0172bc6e-cd47-4fc4-bd87-c0a88d062136",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=44.93\ttotal_loss=44.93\ttime=0.0\n",
      "iter=100\tinput_loss=85.56\toutput_loss=1.468\ttotal_loss=10.02\ttime= 1.2\n",
      "iter=200\tinput_loss=93.56\toutput_loss=0.1706\ttotal_loss=9.527\ttime=1.204\n",
      "iter=300\tinput_loss=52.75\toutput_loss=0.1505\ttotal_loss=5.425\ttime=1.201\n",
      "iter=400\tinput_loss=48.0\toutput_loss=0.04736\ttotal_loss=4.847\ttime=1.184\n",
      "iter=500\tinput_loss=41.31\toutput_loss=0.007231\ttotal_loss=4.138\ttime=1.182\n",
      "iter=600\tinput_loss=43.44\toutput_loss=0.05472\ttotal_loss=4.398\ttime=1.185\n",
      "iter=700\tinput_loss=39.62\toutput_loss=0.1141\ttotal_loss=4.077\ttime=1.183\n",
      "iter=F\tinput_loss=34.44\toutput_loss=0.05474\ttotal_loss=3.498\ttime=9.109\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor(5.9989)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def ballpark_loss(y_hat, y_bar, z=1):    \n",
    "    return torch.maximum(torch.nn.MSELoss(reduction='none')(y_hat, y_bar) - z, torch.zeros_like(y_hat)).mean()\n",
    "\n",
    "y_bar = torch.tensor([[7.0]])\n",
    "X_bar = ledidi(bpnet_ctcf, X, y_bar, output_loss=ballpark_loss, verbose=True)\n",
    "\n",
    "y_hat = predict(bpnet_ctcf, X_bar)\n",
    "y_hat.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "327c7697-2ad7-4105-84bc-df2fc148dd3f",
   "metadata": {},
   "source": [
    "Looks like we get a value near 6. This is not particularly surprising because we had a target value of 7 and are giving it slack to be anywhere between 6 and 8. Because we are also penalizing edits, we end up near the lower end of the desired range because that requires making the fewest edits."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ec51527-f5c1-441f-81ce-5af63250b2b1",
   "metadata": {},
   "source": [
    "#### One-Directional Loss\n",
    "\n",
    "Potentially, when specifying your loss, what you really want are designs that are that strong *or more extreme.* Basically, you want to make sure that the designs are at least some minimum strength but if the procedure is able to be better than that, even better! There are a few ways that this can be specified. First, we can just say that we will not penalize designs that are greater than the target value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "e95afc40-abfe-495f-869d-88bbe57bced7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=45.93\ttotal_loss=45.93\ttime=0.0\n",
      "iter=100\tinput_loss=808.9\toutput_loss=0.7674\ttotal_loss=0.7674\ttime=1.184\n",
      "iter=200\tinput_loss=563.3\toutput_loss=0.0157\ttotal_loss=0.0157\ttime=1.181\n",
      "iter=F\tinput_loss=795.7\toutput_loss= 0.0\ttotal_loss= 0.0\ttime=2.884\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor(7.3477)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def one_directional_loss(y_hat, y_bar):\n",
    "    return torch.where(y_hat < y_bar, torch.nn.MSELoss(reduction='none')(y_hat, y_bar), torch.zeros_like(y_hat)).mean()\n",
    "\n",
    "\n",
    "y_bar = torch.tensor([[7.0]])\n",
    "X_bar = ledidi(bpnet_ctcf, X, y_bar, l=0, output_loss=one_directional_loss, verbose=True)\n",
    "\n",
    "y_hat = predict(bpnet_ctcf, X_bar)\n",
    "y_hat.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0d3bf469-9d86-426a-86d2-671e32626495",
   "metadata": {},
   "source": [
    "Another way that we could do this is to offer a reward for being above the desired output that is not as strong as the penalty for being below it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "6f806681-01d4-44c8-85c2-9285b6f6fec8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter=I\tinput_loss=0.0\toutput_loss=45.93\ttotal_loss=45.93\ttime=0.0\n",
      "iter=100\tinput_loss=782.8\toutput_loss=0.03437\ttotal_loss=0.03437\ttime=1.181\n",
      "iter=200\tinput_loss=606.1\toutput_loss=-0.002722\ttotal_loss=-0.002722\ttime=1.178\n",
      "iter=300\tinput_loss=857.7\toutput_loss=-0.8195\ttotal_loss=-0.8195\ttime=1.18\n",
      "iter=400\tinput_loss=971.6\toutput_loss=-1.479\ttotal_loss=-1.479\ttime=1.178\n",
      "iter=500\tinput_loss=1.018e+03\toutput_loss=-2.272\ttotal_loss=-2.272\ttime=1.179\n",
      "iter=600\tinput_loss=1.055e+03\toutput_loss=-2.686\ttotal_loss=-2.686\ttime=1.177\n",
      "iter=700\tinput_loss=1.092e+03\toutput_loss=-3.353\ttotal_loss=-3.353\ttime=1.179\n",
      "iter=800\tinput_loss=1.117e+03\toutput_loss=-3.812\ttotal_loss=-3.812\ttime=1.178\n",
      "iter=900\tinput_loss=1.145e+03\toutput_loss=-4.589\ttotal_loss=-4.589\ttime=1.178\n",
      "iter=1000\tinput_loss=1.16e+03\toutput_loss=-5.064\ttotal_loss=-5.064\ttime=1.18\n",
      "iter=F\tinput_loss=1.159e+03\toutput_loss=-5.171\ttotal_loss=-5.171\ttime=11.79\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor(14.1902)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def one_directional_loss(y_hat, y_bar):\n",
    "    mse = torch.nn.MSELoss(reduction='none')(y_hat, y_bar)\n",
    "    return torch.where(y_hat < y_bar, mse, -0.1 * mse).mean()\n",
    "\n",
    "\n",
    "y_bar = torch.tensor([[7.0]])\n",
    "X_bar = ledidi(bpnet_ctcf, X, y_bar, l=0, output_loss=one_directional_loss, verbose=True)\n",
    "\n",
    "y_hat = predict(bpnet_ctcf, X_bar)\n",
    "y_hat.mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "392c6905-25ae-4850-bfbe-5a6369103c70",
   "metadata": {},
   "source": [
    "In both examples, we set lambda to 0 so that we can see the true effects, but balancing lambda with the other terms of the loss will be important for design in practice. Regardless, this is just another example of how creative one can be when coming up with custom losses."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}