The OFFICIAL repository of Basenji is at calico/basenji.
[!TIP]
The MultiMolecule team has confirmed that the provided model and checkpoints are producing the same intermediate representations as the original implementation.
The team releasing Basenji did not write this model card for this model so this model card has been written by the MultiMolecule team.
Model Details
Basenji is a deep convolutional neural network trained to predict genomic regulatory activity from long DNA sequences. It consumes a long DNA window (~131 kb), passes it through a convolution + pooling stem that downsamples the sequence, and then through a tower of dilated residual convolutional blocks that expand the receptive field. A pointwise output head predicts a vector of genomic coverage tracks for each output bin. Because the stem downsamples the input, the prediction is binned: the output has shape (batch_size, num_bins, num_tracks) where each bin summarizes 128 bp of sequence and num_tracks is the number of genomic coverage experiments.
Model Specification
Input Length
Bin Size
Output Bins
Hidden Size
Dilated Blocks
Num Labels
Num Parameters (M)
FLOPs (G)
MACs (G)
Max Num Tokens
131,072
128
896
768
11
5,313
30.09
234.85
117.19
131,072
FLOPs and MACs are measured on the canonical 131,072 bp Basenji input window.
The binned positional axis is treated as the "token" axis: each output position corresponds to one
genomic bin rather than a single nucleotide.
Interface
Input length: fixed 131,072 bp DNA window
Output binning: 128 bp per output bin; 896 output bins per window (after Cropping1D(64) on each side)
Output: raw pre-softplus logits of shape (batch_size, num_bins, num_tracks); use postprocess for non-negative coverage tracks
Training Details
Basenji was trained to predict genomic coverage tracks (DNase-seq, ATAC-seq, ChIP-seq and CAGE) from
the human and mouse reference genomes.
Training Data
The model was trained on a large compendium of functional genomics experiments aligned to the human
(hg38) and mouse (mm10) reference genomes. The genome was divided into overlapping windows; for each
window the per-128-bp coverage of every experiment served as the regression target.
Training Procedure
Pre-training
The model was trained to minimize a Poisson regression loss between predicted and observed coverage.
Citation
bibtex
1@article{kelley2018sequential,
2 author = {Kelley, David R. and Reshef, Yakir A. and Bileschi, Maxwell and Belanger, David and McLean, Cory Y. and Snoek, Jasper},
3 title = {Sequential regulatory activity prediction across chromosomes with deep convolutional and recurrent neural networks},
4 journal = {Genome Research},
5 year = 2018,
6 volume = 28,
7 number = 5,
8 pages = {739--750},
9 doi = {10.1101/gr.227819.117},
10 publisher = {Cold Spring Harbor Laboratory}
11}
[!NOTE]
The artifacts distributed in this repository are part of the MultiMolecule project.
If MultiMolecule supports your research, please cite the MultiMolecule project as follows:
bibtex
1@software{chen_2024_12638419,
2 author = {Chen, Zhiyuan and Zhu, Sophia Y.},
3 title = {MultiMolecule},
4 doi = {10.5281/zenodo.12638419},
5 publisher = {Zenodo},
6 url = {https://doi.org/10.5281/zenodo.12638419},
7 year = 2024,
8 month = may,
9 day = 4
10}
Contact
Please use GitHub issues of MultiMolecule for any questions or comments on the model card.
Please contact the authors of the Basenji paper for questions or comments on the paper/model.
