AI Insight
Researchers compared different machine learning models for predicting gene expression levels from chromatin accessibility data using 126,604 single-cell profiles from 19 human studies. The scButterfly-B model, based on variational autoencoders, consistently outperformed transformer-based models including Omic Translator and Cisformer on key accuracy metrics, while requiring 3.8 times less training time. The performance gap was particularly notable for per-gene correlations, where transformers lagged by 0.10-0.24 points, though they came closer on within-cell profile predictions.
Why it matters
This standardized benchmark provides the first rigorous comparison of computational models that link chromatin accessibility to gene expression, which is crucial for understanding gene regulation. The findings suggest that newer transformer architectures do not always outperform traditional autoencoder approaches for this biological task, which has implications for choosing appropriate models in genomics research and potentially reducing computational costs.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Abstract: Models that predict RNA abundance from chromatin accessibility are rarely compared under the same features, targets and metrics. We built one benchmark on 126,604 paired single-nucleus profiles from 19 human studies with a shared vocabulary of 50,000 peaks and 3,185 genes, ten 5×2 cross-validation folds and one scoring target, and compared seven implemented pipelines and two reference predictors across 70 trained runs; each pipeline keeps its own input normalisation and native training target. The scButterfly-B pipeline, built on dual variational autoencoders, outperformed every transformer pipeline, including an autoregressive Perceiver model (Omic Translator) and a support-free reformulation of Cisformer, on per-gene and within-cell Spearman correlation in all sixty paired fold comparisons, with 3.8 times shorter training. Transformer pipelines come within 0.02-0.04 on within-cell profile shape but lag by 0.10-0.24 on per-gene correlation. On two withheld cell types per-gene correlation is low; controls show the podocyte values are typical of within-type correlation for seen cell types of the same size, whereas Paneth values match a cell-agnostic shuffle for every Omic Translator arm. Within the transformer family, prespecified ablations show the Perceiver encoder was worth +0.0439 in within-donor per-gene correlation over full self-attention, a chromosome-aware genomic attention prior reduced it by 0.0096 (exact donor-level sign-flip test, Holm-adjusted p = 1.2e-4), and decoder output order had no detectable effect. A training-mean profile reaches 0.314 within-cell correlation with no input and a constant at the true expressed-gene positions reaches 0.996 per-gene correlation, so oracle-support protocols are reported only as diagnostics. Code, results and an audited release are provided.