Borzoi

Borzoi — DNA → Functional Track Prediction

Prerequisites

| Requirement | Minimum | Recommended |

| ----------- | ------- | ----------- |

| Python | 3.10+ | 3.11 |

| CUDA | 12.1+ | 12.4+ |

| GPU VRAM | 16 GB | 24 GB+ |

How to run

from borzoi_pytorch import Borzoi

model = Borzoi.from_pretrained("johahi/borzoi-replicate-0").cuda().eval()
# input: (batch, 4, 524288) one-hot DNA  → output: (batch, tracks, 6144) bins

Borzoi consumes ~524 kb one-hot windows and emits binned predictions across

7,611 human tracks (the separate 2,608-track mouse head is off by default;

enable via enable_mouse_head=True and select with

forward(..., is_human=False)). For variant scoring, run ref/alt windows

centred on the variant and compare per-track output.

Output format

(B, T, L) tensor — T tracks × L 32-bp bins. Track metadata (assay,

biosample) is in borzoi_pytorch.pytorch_borzoi_model.TRACKS_DF (or model.tracks_df when using the AnnotatedBorzoi subclass) — the base Borzoi model has no targets attribute.

Remote compute

Needs ≥24 GB VRAM and either pre-cached HF weights or egress to

huggingface.co. Use a selected and probed ssh:<alias> context and load

remote-compute-ssh. Confirm borzoi-pytorch and the cache location, then

submit a self-contained runner with run_in_context:

{
  "context_id": "ssh:gpu-box",
  "title": "Borzoi prediction for one locus",
  "command": "source ~/miniforge3/etc/profile.d/conda.sh && conda activate borzoi && HF_HOME=/srv/model-cache python borzoi_run.py --output /home/me/wisp-results/borzoi/tracks.npz",
  "timeout_secs": 1800,
  "input_paths": ["runs/borzoi_run.py"],
  "output_specs": [
    {
      "glob": "ssh://gpu-box/home/me/wisp-results/borzoi/tracks.npz",
      "kind": "npz",
      "residency": "remote"
    }
  ]
}

Replace context, environment, cache, and output paths with discovered values.

Call monitor_run once to wait, get_run once for a snapshot, or cancel_run

to stop.

Troubleshooting

| Symptom | Cause | Fix |

| ------------------------------ | ------------------------ | ------------------------------------ |

| module has no __version__ | Package exposes no attr | Use importlib.metadata.version("borzoi-pytorch") |

| Shape mismatch on input | Wrong window length | Pad/crop to 524288 bp (fixed; not exposed as a model attribute) |

Next: combine track deltas with evo2 likelihood deltas for a

two-axis variant prioritisation.