Enformer Deepmind Tools#

Configuration File: remote_tools/enformer_deepmind_tools.json Tool Type: Remote Tools Count: 1

This page contains all tools defined in the enformer_deepmind_tools.json configuration file.

Available Tools#

run_enformer_deepmind_predict (Type: RemoteTool)#

Predict genomic tracks from DNA sequence with the OFFICIAL DeepMind Enformer release (the TF-Hub …

run_enformer_deepmind_predict tool specification

Tool Information:

  • Name: run_enformer_deepmind_predict

  • Type: RemoteTool

  • Description: Predict genomic tracks from DNA sequence with the OFFICIAL DeepMind Enformer release (the TF-Hub SavedModel deepmind/enformer/1), byte-pinned and verified. Returns the centre-bin value across every track, plus the full 896-bin profile of any track named. Use this when published Enformer numbers must reproduce; use run_enformer_predict (the enformer-pytorch port, GPU) when throughput matters more than matching the authors’ artifact. One window costs about three minutes: this release has no GPU-capable build for this machine’s architecture and runs on the CPU.

Parameters:

  • sequence (string) (required) DNA sequence (A, C, G, T, N). Centre-cropped or N-padded to 393,216 bases, of which the model reads the central 196,608. The cost does not depend on the length you send.

  • organism (string) (optional) Which head to read: ‘human’ has 5,313 tracks, ‘mouse’ 1,643.

  • tracks (array) (optional) Track indices whose full 896-bin profile to return, at most 32. The centre bin of every track is returned regardless.

Example Usage:

query = {
    "name": "run_enformer_deepmind_predict",
    "arguments": {
        "sequence": "example_value"
    }
}
result = tu.run(query)