Two models, in the browser

Shorkie Lab

Shorkie predicts what a yeast sequence expresses. Shorkie_LM predicts what belongs at each base. Same encoder, different question — and the same drawing means opposite things in each.

Both pages run the real published models — the checkpoints from Chao et al. 2025, converted and executed in your browser, not re-creations or heuristics. Both work over the same fourteen S. cerevisiae windows with the same curated annotation, so a result on one page can be read directly against the other.

The same trunk, a different decoder

The two models share their entire encoder — an 11 bp convolutional stem, seven residual blocks that pool 16,384 positions down to 128, and eight transformer layers at that bottleneck. They diverge only in what they do on the way back up, and the divergence is not arbitrary: each decoder resolves exactly as finely as its output is defined.

ShorkieShorkie_LM
stem · residual tower · transformeridentical — the shared trunk
U-Net decoder blocks3  (128 → 1,024, cropped to 896)7  (128 → 16,384, no crop)
head5,215 units, softplus4 units, softmax
output896 bins × 5,215 tracks16,384 positions × 4 bases
resolution16 bp — a coverage bin1 bp — a base
trainingfine-tuned, lr 2e-5, 8 foldsmasked LM, 15% mask rate, 1 fold

Shorkie stops at 16 bp because coverage is a 16 bp quantity; the LM upsamples all the way back because a base is what it predicts. Everything else in the two networks is the same weights-shaped object trained toward a different loss.

Which page answers which question

“Which bases drive this gene's expression?”
Shorkie. Attribution is a statement about a prediction, and only the expression model makes one.
“Is this stretch of sequence under constraint?”
Shorkie_LM. Low entropy in the predicted distribution means the model finds the base highly determined by its context — the model's own analogue of conservation, computed without an alignment.
“Is that peak a real binding site?”
Either — both pages carry the same curated annotation layer from SGD, the Harbison/MacIsaac regulatory code, ORegAnno and JASPAR, drawn with its evidence tier.
“What did pretraining buy?”
Neither, yet. That needs the published Shorkie_Random_Init ablation, which is not on the site.

Connected resources

Source, weights and documentation: the Shorkie repository and its documentation site. The paper is Chao et al. 2025.