{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "cell-01",
   "metadata": {},
   "source": [
    "# Finetuning AlphaGenome on ATAC-seq with linear probing\n",
    "\n",
    "This notebook demonstrates how to finetune AlphaGenome on GM12878 chromatin accessibility\n",
    "(ATAC-seq) data using linear probing, where the trunk is frozen and only a new head is trained.\n",
    "\n",
    "You will need to run this notebook from within a virtual environment with the `dev` dependencies for `alphagenome-pytorch`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "20fb316f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f95b2d39",
   "metadata": {},
   "source": [
    "## 1. Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "cell-02",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Imports OK\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import csv\n",
    "from datetime import datetime\n",
    "from pathlib import Path\n",
    "\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import torch\n",
    "from scipy.stats import pearsonr\n",
    "from tqdm import tqdm\n",
    "\n",
    "from alphagenome_pytorch import AlphaGenome\n",
    "from alphagenome_pytorch.extensions.finetuning.checkpointing import (\n",
    "    save_checkpoint,\n",
    "    load_checkpoint,\n",
    ")\n",
    "from alphagenome_pytorch.extensions.finetuning.datasets import (\n",
    "    CachedGenome,\n",
    "    GenomicDataset,\n",
    "    compute_track_means,\n",
    ")\n",
    "from alphagenome_pytorch.extensions.finetuning.heads import create_finetuning_head\n",
    "from alphagenome_pytorch.extensions.finetuning.training import (\n",
    "    collate_genomic,\n",
    "    compute_finetuning_loss,\n",
    "    create_lr_scheduler,\n",
    "    train_epoch,\n",
    "    validate,\n",
    ")\n",
    "from alphagenome_pytorch.extensions.finetuning.transfer import load_trunk\n",
    "\n",
    "matplotlib.rcParams['figure.dpi'] = 100\n",
    "print('Imports OK')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-03",
   "metadata": {},
   "source": [
    "## 2. Configuration\n",
    "\n",
    "Define hyperparameters and files needed for finetuning. The following hyperparmameters run on an A100.\n",
    "\n",
    "### Input data:\n",
    "- `GENOME_FASTA`: A reference genome FASTA file\n",
    "- `BIGWIG_FILES`: A series of signal BigWig files to train on.\n",
    "- `TRAIN_BED`: Genomic regions for training. Here, we use regions of size 131,072 base pairs from segments specified in the Borzoi FOLD_0 training set.\n",
    "- `VAL_BED`: Genomic regions for validation. Here, we use regions of size 131,072 base pairs from segments specified in the Borzoi FOLD_0 validation set.\n",
    "- `PRETRAINED_WEIGHTS`: Pretrained weights from the AlphaGenome model. We use a conversion from the FOLD_0 model.\n",
    "\n",
    "### Model settings\n",
    "- `MODALITY`: Prediction modality.\n",
    "- `RESOLUTIONS`: Can be `(1,)`, `(128,)`, `(1, 128)`\n",
    "- `SEQUENCE_LENGTH`: Length of input DNA sequences. We train on 2**17 sequences, can be up to 2\\*\\*20."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "cell-04",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Configuration:\n",
      "  MODALITY        : atac\n",
      "  RESOLUTIONS     : (1,)\n",
      "  SEQUENCE_LENGTH : 131,072 bp\n",
      "  BATCH_SIZE      : 4\n",
      "  EPOCHS          : 10\n",
      "  LR              : 0.0001\n",
      "  WARMUP_STEPS    : 500\n",
      "  OUTPUT_DIR      : finetuning_output/20260302_061328\n",
      "  COMPILE         : True\n"
     ]
    }
   ],
   "source": [
    "# ── Data paths ────────────────────────────────────────────────────────────────\n",
    "GENOME_FASTA = 'finetune_files/hg38.fa'\n",
    "BIGWIG_FILES = ['finetune_files/unstranded.bw']\n",
    "TRAIN_BED = 'finetune_files/ag_regions/fold_0/train.bed'\n",
    "VAL_BED = 'finetune_files/ag_regions/fold_0/valid.bed'\n",
    "PRETRAINED_WEIGHTS = 'finetune_files/weights/fold_0_weights.pth'\n",
    "\n",
    "# ── Model ─────────────────────────────────────────────────────────────────────\n",
    "MODALITY = 'atac'\n",
    "RESOLUTIONS = (1,)  # 1bp only (no 128bp head)\n",
    "SEQUENCE_LENGTH = 131_072  # 131kb windows\n",
    "\n",
    "# ── Training Hyperparameters ───────────────────────────────────────────────────\n",
    "BATCH_SIZE = 4\n",
    "EPOCHS = 10\n",
    "LR = 1e-4\n",
    "WEIGHT_DECAY = 0.1\n",
    "WARMUP_STEPS = 500\n",
    "POSITIONAL_WEIGHT = 5.0\n",
    "NUM_WORKERS = 1\n",
    "COMPILE = True\n",
    "\n",
    "# ── Output ────────────────────────────────────────────────────────────────────\n",
    "run_id = datetime.now().strftime('%Y%m%d_%H%M%S')\n",
    "OUTPUT_DIR = Path('finetuning_output') / run_id\n",
    "OUTPUT_DIR.mkdir(parents=True, exist_ok=True)\n",
    "\n",
    "print('Configuration:')\n",
    "print(f'  MODALITY        : {MODALITY}')\n",
    "print(f'  RESOLUTIONS     : {RESOLUTIONS}')\n",
    "print(f'  SEQUENCE_LENGTH : {SEQUENCE_LENGTH:,} bp')\n",
    "print(f'  BATCH_SIZE      : {BATCH_SIZE}')\n",
    "print(f'  EPOCHS          : {EPOCHS}')\n",
    "print(f'  LR              : {LR}')\n",
    "print(f'  WARMUP_STEPS    : {WARMUP_STEPS}')\n",
    "print(f'  OUTPUT_DIR      : {OUTPUT_DIR}')\n",
    "print(f'  COMPILE         : {COMPILE}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "cell-05",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda\n",
      "Resolution weights: {1: 1.0}\n"
     ]
    }
   ],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "print(f'Device: {device}')\n",
    "# Resolution weights — all resolutions equally weighted\n",
    "resolution_weights = {res: 1.0 for res in RESOLUTIONS}\n",
    "print(f'Resolution weights: {resolution_weights}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-06",
   "metadata": {},
   "source": [
    "## 3. Data Loading\n",
    "\n",
    "We have 41,694 training and 7,090 validation genomic windows, each 2**17 kb. The BED\n",
    "file defines windows; `GenomicDataset` expands each interval to exactly `sequence_length` from\n",
    "its center.\n",
    "\n",
    "Each sample returns:\n",
    "- **`sequence`**: one-hot encoded DNA, shape `(131072, 4)`\n",
    "- **`targets_dict[1]`**: raw ATAC-seq signal at 1bp resolution, shape `(131072, 1)`\n",
    "\n",
    "`CachedGenome` pre-loads the entire hg38 genome (~12 GB) into RAM as one-hot arrays.\n",
    "Pass the same `CachedGenome` instance to both datasets to share the cache and halve memory usage."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "cell-07",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading genome into memory (hg38, ~12 GB)...\n",
      "  (This takes 3-5 minutes on first run)\n",
      "CachedGenome: Loading genome from finetune_files/hg38.fa...\n",
      "CachedGenome: Loaded 455 chromosomes (12837.1 MB)\n",
      "\n",
      "Creating train/val datasets...\n",
      "\n",
      "Train: 41,694  Val: 6,461\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/scratch/m000097/abuen/personal/alphagenome-pytorch/src/alphagenome_pytorch/extensions/finetuning/datasets.py:547: UserWarning: 41694 intervals were larger than sequence_length=131072 and were centered and truncated, which may lose important flanking regions.\n",
      "  warnings.warn(\n",
      "/scratch/m000097/abuen/personal/alphagenome-pytorch/src/alphagenome_pytorch/extensions/finetuning/datasets.py:547: UserWarning: 6461 intervals were larger than sequence_length=131072 and were centered and truncated, which may lose important flanking regions.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "print('Loading genome into memory (hg38, ~12 GB)...')\n",
    "print('  (This takes 3-5 minutes on first run)')\n",
    "shared_genome = CachedGenome(GENOME_FASTA)\n",
    "\n",
    "print('\\nCreating train/val datasets...')\n",
    "# Pass the CachedGenome directly as genome_fasta to share the cache\n",
    "train_dataset = GenomicDataset(\n",
    "    genome_fasta=shared_genome,\n",
    "    bigwig_files=BIGWIG_FILES,\n",
    "    bed_file=TRAIN_BED,\n",
    "    resolutions=RESOLUTIONS,\n",
    "    sequence_length=SEQUENCE_LENGTH,\n",
    ")\n",
    "val_dataset = GenomicDataset(\n",
    "    genome_fasta=shared_genome,\n",
    "    bigwig_files=BIGWIG_FILES,\n",
    "    bed_file=VAL_BED,\n",
    "    resolutions=RESOLUTIONS,\n",
    "    sequence_length=SEQUENCE_LENGTH,\n",
    ")\n",
    "print(f'\\nTrain: {len(train_dataset):,}  Val: {len(val_dataset):,}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "cell-09",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train loader: 10,424 batches/epoch\n",
      "Val loader  : 1,616 batches\n"
     ]
    }
   ],
   "source": [
    "train_loader = torch.utils.data.DataLoader(\n",
    "    train_dataset,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    shuffle=True,\n",
    "    num_workers=NUM_WORKERS,\n",
    "    pin_memory=True,\n",
    "    collate_fn=collate_genomic,\n",
    "    prefetch_factor=2 if NUM_WORKERS > 0 else None,\n",
    "    persistent_workers=NUM_WORKERS > 0,\n",
    ")\n",
    "val_loader = torch.utils.data.DataLoader(\n",
    "    val_dataset,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    shuffle=False,\n",
    "    num_workers=NUM_WORKERS,\n",
    "    pin_memory=True,\n",
    "    collate_fn=collate_genomic,\n",
    "    prefetch_factor=2 if NUM_WORKERS > 0 else None,\n",
    "    persistent_workers=NUM_WORKERS > 0,\n",
    ")\n",
    "\n",
    "print(f'Train loader: {len(train_loader):,} batches/epoch')\n",
    "print(f'Val loader  : {len(val_loader):,} batches')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-10",
   "metadata": {},
   "source": [
    "## 4. Model Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cell-11",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading pretrained AlphaGenome weights from finetune_files/weights/fold_0_weights.pth...\n",
      "Backbone frozen.\n",
      "  Backbone parameters: 450,452,613\n"
     ]
    }
   ],
   "source": [
    "print(f'Loading pretrained AlphaGenome weights from {PRETRAINED_WEIGHTS}...')\n",
    "model = AlphaGenome()\n",
    "model = load_trunk(model, PRETRAINED_WEIGHTS, exclude_heads=True)\n",
    "\n",
    "# Freeze backbone — only the new head will be trained\n",
    "for param in model.parameters():\n",
    "    param.requires_grad = False\n",
    "\n",
    "n_backbone = sum(p.numel() for p in model.parameters())\n",
    "print(f'Backbone frozen.')\n",
    "print(f'  Backbone parameters: {n_backbone:,}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "cell-12",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Computing track means from training data...\n",
      "Computed nonzero_mean per track: [2.24096773]\n",
      "Compiling model (first run will be slow)...\n",
      "\n",
      "Trainable: 1,538 / 448,092,807 (0.0003%)\n"
     ]
    }
   ],
   "source": [
    "print('Computing track means from training data...')\n",
    "track_means = compute_track_means(\n",
    "    BIGWIG_FILES,\n",
    "    TRAIN_BED,\n",
    "    sequence_length=SEQUENCE_LENGTH,\n",
    "    max_samples=1000,\n",
    ")\n",
    "\n",
    "# Create finetuning head and attach to model\n",
    "head = create_finetuning_head(\n",
    "    assay_type=MODALITY,\n",
    "    n_tracks=len(BIGWIG_FILES),\n",
    "    resolutions=RESOLUTIONS,\n",
    "    num_organisms=1,\n",
    "    track_means=track_means,\n",
    ")\n",
    "model.heads[MODALITY] = head\n",
    "model = model.to(device)\n",
    "head = model.heads[MODALITY]  # update reference after .to()\n",
    "\n",
    "if COMPILE:\n",
    "    print(\"Compiling model (first run will be slow)...\")\n",
    "    model = torch.compile(model)\n",
    "\n",
    "n_trainable = sum(p.numel() for p in head.parameters())\n",
    "n_total = sum(p.numel() for p in model.parameters())\n",
    "print(f'\\nTrainable: {n_trainable:,} / {n_total:,} ({100 * n_trainable / n_total:.4f}%)')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-13",
   "metadata": {},
   "source": [
    "## 5. Training Setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "cell-14",
   "metadata": {},
   "outputs": [],
   "source": [
    "optimizer = torch.optim.AdamW(\n",
    "    head.parameters(),\n",
    "    lr=LR,\n",
    "    weight_decay=WEIGHT_DECAY,\n",
    ")\n",
    "\n",
    "steps_per_epoch = len(train_loader)\n",
    "total_steps = EPOCHS * steps_per_epoch\n",
    "scheduler = create_lr_scheduler(optimizer, WARMUP_STEPS, total_steps, schedule='cosine')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-15",
   "metadata": {},
   "source": [
    "AlphaGenome uses a loss with two parts:\n",
    "\n",
    "1. **Positional (multinomial) loss** (`loss_positional`). Upweighted by `positional_weight=5.0`.\n",
    "2. **Count (Poisson) loss** (`loss_count`).\n",
    "\n",
    "The multinomial loss is computed across `NUM_SEGMENTS=8` equal segments for numerical stability.\n",
    "The sequence length (131,072 bp) is divided into 8 segments of 16,384 bp each.\n",
    "\n",
    "We use `compute_finetuning_loss` from the library, which handles this internally.\n",
    "The library also provides `train_epoch` and `validate` functions that implement the full\n",
    "training loop with gradient accumulation, AMP, and progress bars."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-17",
   "metadata": {},
   "source": [
    "## 6. Training Loop"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cell-19",
   "metadata": {},
   "outputs": [],
   "source": [
    "epoch_log_rows = []\n",
    "best_val_loss = float('inf')\n",
    "\n",
    "print(f'Starting training: {EPOCHS} epochs, batch size {BATCH_SIZE}')\n",
    "print(f'Outputs → {OUTPUT_DIR}')\n",
    "print('=' * 60)\n",
    "\n",
    "for epoch in range(1, EPOCHS + 1):\n",
    "    train_loss = train_epoch(\n",
    "        model=model,\n",
    "        head=head,\n",
    "        train_loader=train_loader,\n",
    "        optimizer=optimizer,\n",
    "        scheduler=scheduler,\n",
    "        device=device,\n",
    "        resolution_weights=resolution_weights,\n",
    "        positional_weight=POSITIONAL_WEIGHT,\n",
    "        epoch=epoch,\n",
    "        log_every=50,\n",
    "    )\n",
    "\n",
    "    val_loss = validate(\n",
    "        model=model,\n",
    "        head=head,\n",
    "        val_loader=val_loader,\n",
    "        device=device,\n",
    "        resolution_weights=resolution_weights,\n",
    "        positional_weight=POSITIONAL_WEIGHT,\n",
    "    )\n",
    "\n",
    "    current_lr = scheduler.get_last_lr()[0]\n",
    "    is_best = val_loss < best_val_loss\n",
    "\n",
    "    star = '★ best' if is_best else ''\n",
    "    print(f'\\nEpoch {epoch:2d} | train={train_loss:.4f}  val={val_loss:.4f}'\n",
    "          f'  lr={current_lr:.2e}  {star}')\n",
    "\n",
    "    epoch_row = {\n",
    "        'epoch': epoch,\n",
    "        'train_loss': train_loss,\n",
    "        'val_loss': val_loss,\n",
    "        'learning_rate': current_lr,\n",
    "        'is_best': is_best,\n",
    "        'timestamp': datetime.now().isoformat(),\n",
    "    }\n",
    "    epoch_log_rows.append(epoch_row)\n",
    "\n",
    "    if is_best:\n",
    "        best_val_loss = val_loss\n",
    "        save_checkpoint(\n",
    "            path=OUTPUT_DIR / 'best_model.pth',\n",
    "            epoch=epoch,\n",
    "            model=model,\n",
    "            optimizer=optimizer,\n",
    "            scheduler=scheduler,\n",
    "            val_loss=val_loss,\n",
    "            best_val_loss=best_val_loss,\n",
    "            track_names=[Path(BIGWIG_FILES[0]).stem],\n",
    "            modality=MODALITY,\n",
    "            resolutions=RESOLUTIONS,\n",
    "        )\n",
    "        print(f'  -> Best model saved (val={val_loss:.4f})')\n",
    "\n",
    "    save_checkpoint(\n",
    "        path=OUTPUT_DIR / f'checkpoint_epoch{epoch}.pth',\n",
    "        epoch=epoch,\n",
    "        model=model,\n",
    "        optimizer=optimizer,\n",
    "        scheduler=scheduler,\n",
    "        val_loss=val_loss,\n",
    "        best_val_loss=best_val_loss,\n",
    "        track_names=[Path(BIGWIG_FILES[0]).stem],\n",
    "        modality=MODALITY,\n",
    "        resolutions=RESOLUTIONS,\n",
    "    )\n",
    "\n",
    "# ── Write CSV ──────────────────────────────────────────────────────────────────\n",
    "with open(OUTPUT_DIR / 'epoch_log.csv', 'w', newline='') as f:\n",
    "    if epoch_log_rows:\n",
    "        writer = csv.DictWriter(f, fieldnames=list(epoch_log_rows[0].keys()))\n",
    "        writer.writeheader()\n",
    "        writer.writerows(epoch_log_rows)\n",
    "\n",
    "print(f'\\n{\"=\"*60}')\n",
    "print(f'Training complete!')\n",
    "print(f'Best val loss : {best_val_loss:.4f}')\n",
    "print(f'Output        : {OUTPUT_DIR}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-20",
   "metadata": {},
   "source": [
    "## 7. Results Visualization\n",
    "\n",
    "Training converged quickly over 10 epochs:\n",
    "\n",
    "- **Loss decomposition**: the total multinomial loss is dominated by the positional\n",
    "  component, meaning peak shape is the harder prediction task. The count component is small — predicting total read depth from sequence is easier.\n",
    "- **Train/val gap**: minimal, indicating no significant overfitting despite only 1,538 trained parameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "7572c1d9-d350-4945-b543-8ce01cfb354f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# uncomment for demo run results\n",
    "# OUTPUT_DIR = Path('finetuning_output') / \"20260301_153530\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "cell-21",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== Training Summary ===\n",
      "Epochs completed  : 10\n",
      "Best val loss     : 7.722594  (epoch 6)\n",
      "Final train loss  : 7.982453\n",
      "Final val loss    : 7.722741\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "epoch_df = pd.read_csv(OUTPUT_DIR / 'epoch_log.csv')\n",
    "\n",
    "best_row = epoch_df.loc[epoch_df['val_loss'].idxmin()]\n",
    "final_row = epoch_df.iloc[-1]\n",
    "\n",
    "print('=== Training Summary ===')\n",
    "print(f'Epochs completed  : {len(epoch_df)}')\n",
    "print(f'Best val loss     : {best_row[\"val_loss\"]:.6f}  (epoch {int(best_row[\"epoch\"])})')\n",
    "print(f'Final train loss  : {final_row[\"train_loss\"]:.6f}')\n",
    "print(f'Final val loss    : {final_row[\"val_loss\"]:.6f}')\n",
    "\n",
    "fig, axes = plt.subplots(1, 2, figsize=(12, 4))\n",
    "fig.suptitle(\n",
    "    f'Finetuning run: {OUTPUT_DIR.name}  (GM12878 ATAC-seq)',\n",
    "    fontsize=13, fontweight='bold'\n",
    ")\n",
    "\n",
    "# ── Val loss per epoch ─────────────────────────────────────────────────────────\n",
    "ax = axes[0]\n",
    "ax.plot(epoch_df['epoch'], epoch_df['val_loss'], marker='s', color='tomato', label='Val loss')\n",
    "ax.axvline(int(best_row['epoch']), color='gray', linestyle='--', linewidth=0.8, alpha=0.7)\n",
    "ax.set_xlabel('Epoch'); ax.set_ylabel('Loss')\n",
    "ax.set_title('Validation Loss (per epoch)'); ax.legend(); ax.grid(True, alpha=0.3)\n",
    "ax.set_xticks(epoch_df['epoch'])\n",
    "\n",
    "# ── Train loss per epoch ───────────────────────────────────────────────────────\n",
    "ax = axes[1]\n",
    "ax.plot(epoch_df['epoch'], epoch_df['train_loss'], marker='o', color='steelblue', label='Train loss')\n",
    "ax.set_xlabel('Epoch'); ax.set_ylabel('Loss')\n",
    "ax.set_title('Training Loss (per epoch)'); ax.legend(); ax.grid(True, alpha=0.3)\n",
    "ax.set_xticks(epoch_df['epoch'])\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "cell-22",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch-level summary:\n"
     ]
    },
    {
     "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>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>val_loss</th>\n",
       "      <th>learning_rate</th>\n",
       "      <th>is_best</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>8.144648</td>\n",
       "      <td>7.730252</td>\n",
       "      <td>0.000098</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>7.989362</td>\n",
       "      <td>7.725195</td>\n",
       "      <td>0.000091</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>7.985308</td>\n",
       "      <td>7.724322</td>\n",
       "      <td>0.000080</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>7.984085</td>\n",
       "      <td>7.723439</td>\n",
       "      <td>0.000066</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>7.983350</td>\n",
       "      <td>7.723115</td>\n",
       "      <td>0.000050</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>7.983046</td>\n",
       "      <td>7.722594</td>\n",
       "      <td>0.000035</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>7.982787</td>\n",
       "      <td>7.722927</td>\n",
       "      <td>0.000021</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>7.982798</td>\n",
       "      <td>7.722734</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>7.983119</td>\n",
       "      <td>7.722805</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>7.982453</td>\n",
       "      <td>7.722741</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   epoch  train_loss  val_loss  learning_rate  is_best\n",
       "0      1    8.144648  7.730252       0.000098     True\n",
       "1      2    7.989362  7.725195       0.000091     True\n",
       "2      3    7.985308  7.724322       0.000080     True\n",
       "3      4    7.984085  7.723439       0.000066     True\n",
       "4      5    7.983350  7.723115       0.000050     True\n",
       "5      6    7.983046  7.722594       0.000035     True\n",
       "6      7    7.982787  7.722927       0.000021    False\n",
       "7      8    7.982798  7.722734       0.000010    False\n",
       "8      9    7.983119  7.722805       0.000002    False\n",
       "9     10    7.982453  7.722741       0.000000    False"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "display_cols = ['epoch', 'train_loss', 'val_loss', 'learning_rate', 'is_best']\n",
    "\n",
    "print('Epoch-level summary:')\n",
    "epoch_df[display_cols].round(6)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cell-23",
   "metadata": {},
   "source": [
    "## 8. Evaluation: Comparing Fine-tuned to Baseline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "5d3998f7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Baseline (untrained) head created.\n",
      "  Parameters : 1,538  (randomly initialized, no training)\n"
     ]
    }
   ],
   "source": [
    "# Create a fresh head with the same architecture but random (untrained) weights\n",
    "baseline_head = create_finetuning_head(\n",
    "    assay_type=MODALITY,\n",
    "    n_tracks=len(BIGWIG_FILES),\n",
    "    resolutions=RESOLUTIONS,\n",
    "    num_organisms=1,\n",
    "    track_means=track_means,\n",
    ").to(device)\n",
    "baseline_head.eval()\n",
    "\n",
    "n_baseline = sum(p.numel() for p in baseline_head.parameters())\n",
    "print('Baseline (untrained) head created.')\n",
    "print(f'  Parameters : {n_baseline:,}  (randomly initialized, no training)')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "27092e49",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Backbone embeddings computed for val sample 0:\n",
      "  embeddings_1bp : (1, 131072, 1536)\n",
      "Loading best model from finetuning_output/20260301_153530/best_model.pth...\n"
     ]
    }
   ],
   "source": [
    "# Retrieve a validation sample for the predicted vs. fine-tuned set\n",
    "VAL_IDX = 0\n",
    "val_seq, val_targets = val_dataset[VAL_IDX]\n",
    "seq_input = val_seq.unsqueeze(0).to(device)  # (1, 131072, 4)\n",
    "organism_idx = torch.zeros(1, dtype=torch.long, device=device)\n",
    "\n",
    "# Compute the backbone embeddings using return_embeddings=True\n",
    "model.eval()\n",
    "with torch.no_grad():\n",
    "    with torch.autocast(device_type='cuda', dtype=torch.bfloat16,\n",
    "                        enabled=device.type == 'cuda'):\n",
    "        outputs = model(\n",
    "            seq_input, organism_idx,\n",
    "            return_embeddings=True,\n",
    "        )\n",
    "\n",
    "embeddings_dict = {}\n",
    "if 1 in RESOLUTIONS and outputs.get('embeddings_1bp') is not None:\n",
    "    embeddings_dict[1] = outputs['embeddings_1bp']\n",
    "if 128 in RESOLUTIONS and outputs.get('embeddings_128bp') is not None:\n",
    "    embeddings_dict[128] = outputs['embeddings_128bp']\n",
    "\n",
    "print(f'Backbone embeddings computed for val sample {VAL_IDX}:')\n",
    "for res, emb in embeddings_dict.items():\n",
    "    print(f'  embeddings_{res}bp : {tuple(emb.shape)}')\n",
    "\n",
    "# Baseline predictions (untrained head)\n",
    "with torch.no_grad():\n",
    "    with torch.autocast(device_type='cuda', dtype=torch.bfloat16,\n",
    "                        enabled=device.type == 'cuda'):\n",
    "        baseline_preds = baseline_head(embeddings_dict, organism_idx)\n",
    "\n",
    "# Load best checkpoint and get fine-tuned predictions\n",
    "best_ckpt_path = OUTPUT_DIR / 'best_model.pth'\n",
    "print(f'Loading best model from {best_ckpt_path}...')\n",
    "load_checkpoint(\n",
    "    path=best_ckpt_path,\n",
    "    model=model,\n",
    "    optimizer=optimizer,\n",
    "    scheduler=scheduler,\n",
    ")\n",
    "model.eval()\n",
    "head.eval()\n",
    "\n",
    "with torch.no_grad():\n",
    "    with torch.autocast(device_type='cuda', dtype=torch.bfloat16,\n",
    "                        enabled=device.type == 'cuda'):\n",
    "        finetuned_preds = head(embeddings_dict, organism_idx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "691888f2-7f67-4215-836b-42aad780b646",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running pretrained model inference for GM12878 ATAC track...\n",
      "Warning: Missing keys (not loaded): ['heads.atac.track_means', 'heads.dnase.track_means', 'heads.procap.track_means', 'heads.cage.track_means', 'heads.rna_seq.track_means']...\n"
     ]
    }
   ],
   "source": [
    "# Run inference on pretrained model to get GM12878 predictions\n",
    "from alphagenome.models import dna_output\n",
    "from alphagenome_research.model import dna_model as ag_dna_model\n",
    "from alphagenome_research.model.metadata import metadata as metadata_lib\n",
    "\n",
    "ATAC_ONTOLOGY_CURIE = 'EFO:0002784'  # GM12878\n",
    "metadata = metadata_lib.load(ag_dna_model.Organism.HOMO_SAPIENS)\n",
    "track_masks = metadata_lib.create_track_masks(\n",
    "    metadata,\n",
    "    requested_outputs={dna_output.OutputType.ATAC},\n",
    "    requested_ontologies=[dna_output.ontology.from_curie(ATAC_ONTOLOGY_CURIE)],\n",
    ")\n",
    "atac_mask = track_masks[dna_output.OutputType.ATAC]\n",
    "gm12878_track_indices = np.flatnonzero(atac_mask)\n",
    "ATAC_GM12878_TRACK_IDX = int(gm12878_track_indices[0])\n",
    "\n",
    "print('Running pretrained model inference for GM12878 ATAC track...')\n",
    "pretrained_model = AlphaGenome()\n",
    "pretrained_model = load_trunk(pretrained_model, PRETRAINED_WEIGHTS, exclude_heads=False)\n",
    "pretrained_model = pretrained_model.to(device)\n",
    "pretrained_model.eval()\n",
    "\n",
    "with torch.no_grad():\n",
    "    with torch.autocast(device_type='cuda', dtype=torch.bfloat16, enabled=device.type == 'cuda'):\n",
    "        pretrained_outputs = pretrained_model(\n",
    "            seq_input,\n",
    "            organism_idx,\n",
    "            resolutions=(1,),\n",
    "        )\n",
    "pretrained_atac_gm12878 = pretrained_outputs['atac'][1][:, :, ATAC_GM12878_TRACK_IDX:ATAC_GM12878_TRACK_IDX + 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "bc180189",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2000x400 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize the differences between ground truth, finetuned, baseline, and pretrained\n",
    "from alphagenome.visualization import plot_components\n",
    "from alphagenome.data import genome as jax_genome\n",
    "from alphagenome_pytorch.variant_scoring.visualization_utils import pytorch_to_track_data\n",
    "\n",
    "chrom, start, end = val_dataset._positions_list[VAL_IDX]\n",
    "jax_interval = jax_genome.Interval(chromosome=chrom, start=start, end=end)\n",
    "\n",
    "track_meta = type('M', (), {\n",
    "    'track_name': 'GM12878 ATAC-seq',\n",
    "    'track_strand': '.',\n",
    "    'biosample_name': 'GM12878',\n",
    "    'ontology_curie': 'EFO:0002784',\n",
    "    'assay_title': 'ATAC-seq',\n",
    "    'transcription_factor': '',\n",
    "    'histone_mark': '',\n",
    "})()\n",
    "metadata_list = [track_meta]\n",
    "\n",
    "gt_tdata = pytorch_to_track_data(val_targets[1], metadata_list, jax_interval, resolution=1)\n",
    "bl_tdata = pytorch_to_track_data(baseline_preds[1] - baseline_preds[1].mean(), metadata_list, jax_interval, resolution=1)  # Zero center\n",
    "ft_tdata = pytorch_to_track_data(finetuned_preds[1], metadata_list, jax_interval, resolution=1)\n",
    "pt_tdata = pytorch_to_track_data(pretrained_atac_gm12878, metadata_list, jax_interval, resolution=1)\n",
    "\n",
    "components = [\n",
    "    plot_components.Tracks(\n",
    "        tdata=gt_tdata,\n",
    "        ylabel_template='Ground Truth',\n",
    "    ),\n",
    "    plot_components.Tracks(\n",
    "        tdata=bl_tdata,\n",
    "        ylabel_template='Zero-centered Baseline',\n",
    "    ),\n",
    "    plot_components.Tracks(\n",
    "        tdata=pt_tdata,\n",
    "        ylabel_template='Pretrained',\n",
    "    ),\n",
    "    plot_components.Tracks(\n",
    "        tdata=ft_tdata,\n",
    "        ylabel_template='Finetuned',\n",
    "    ),\n",
    "]\n",
    "\n",
    "plot_interval = jax_interval.resize(40000)\n",
    "\n",
    "plot_components.plot(\n",
    "    components=components,\n",
    "    interval=plot_interval,\n",
    "    title='Predicted chromatin accessibility ATAC for GM12878',\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "4c4fd715",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==================================================\n",
      "Pearson r (ground truth vs untrained baseline):\n",
      "  r = -0.1696\n",
      "Pearson r (ground truth vs pretrained model):\n",
      "  r = 0.5890\n",
      "Pearson r (ground truth vs finetuned head):\n",
      "  r = 0.6228\n",
      "Improvement (finetuned - baseline): +0.7924\n",
      "Improvement (finetuned - pretrained): +0.0337\n",
      "==================================================\n"
     ]
    }
   ],
   "source": [
    "# ── Pearson correlation ────────────────────────────────────────────────────────\n",
    "BIN = 128\n",
    "n_bins = SEQUENCE_LENGTH // BIN\n",
    "\n",
    "# Bin to 128bp for comparison\n",
    "ft_1bp_tensor = finetuned_preds[1][0, :, 0].float().cpu()\n",
    "ft_binned = ft_1bp_tensor.numpy()[:n_bins * BIN].reshape(n_bins, BIN).mean(axis=1)\n",
    "\n",
    "gt_1bp_tensor = val_targets[1][:, 0].float().cpu()\n",
    "gt_binned = gt_1bp_tensor.numpy()[:n_bins * BIN].reshape(n_bins, BIN).mean(axis=1)\n",
    "\n",
    "bl_1bp_tensor = baseline_preds[1][0, :, 0].float().cpu()\n",
    "bl_binned = bl_1bp_tensor.numpy()[:n_bins * BIN].reshape(n_bins, BIN).mean(axis=1)\n",
    "\n",
    "pt_1bp_tensor = pretrained_atac_gm12878[0, :, 0].float().cpu()\n",
    "pt_binned = pt_1bp_tensor.numpy()[:n_bins * BIN].reshape(n_bins, BIN).mean(axis=1)\n",
    "\n",
    "r_finetuned, _ = pearsonr(gt_binned, ft_binned)\n",
    "r_baseline, _ = pearsonr(gt_binned, bl_binned)\n",
    "r_pretrained, _ = pearsonr(gt_binned, pt_binned)\n",
    "\n",
    "print('=' * 50)\n",
    "print('Pearson r (ground truth vs untrained baseline):')\n",
    "print(f'  r = {r_baseline:.4f}')\n",
    "\n",
    "print('Pearson r (ground truth vs pretrained model):')\n",
    "print(f'  r = {r_pretrained:.4f}')\n",
    "\n",
    "print('Pearson r (ground truth vs finetuned head):')\n",
    "print(f'  r = {r_finetuned:.4f}')\n",
    "\n",
    "print(f'Improvement (finetuned - baseline): {r_finetuned - r_baseline:+.4f}')\n",
    "print(f'Improvement (finetuned - pretrained): {r_finetuned - r_pretrained:+.4f}')\n",
    "print('=' * 50)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ag-new",
   "language": "python",
   "name": "ag-new"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}