HiCMamba Component Ablation Results for Hi-C Resolution Enhancement

Ablation results of each component of the HiCMamba deep learning model for enhancing Hi-C contact map resolution. The data, provided by Minghao Yang, is associated with a method that uses a state space model and UNet-based architecture to improve chromatin interaction frequency estimates. The dataset is stored in an XLS file with a size of 5.5 KB.

Use Cases

• Compare the performance impact of the holistic scan block component on Hi-C contact map enhancement.
• Analyze the computational resource reduction achieved by the HiCMamba state space model framework.
• Validate identified 3D genome structures like topologically associating domains (TADs) and loops against associated epigenomic features using the model's output.

Strengths

• Results are directly linked to the published HiCMamba method, which demonstrated performance exceeding state-of-the-art alternatives.
• The dataset is openly available under a CC BY 4.0 license, facilitating reuse and verification.
• The associated source code is publicly accessible on GitHub, allowing for reproducibility and extension of the analysis.

Limitations

• The dataset is very small at 5.5 KB, indicating limited scope and likely containing only summary or key result metrics rather than raw experimental data.
• Specific quantitative details such as row count, column structure, and sample data are unavailable, limiting immediate analytical utility.
• The data is specific to the ablation study of a single model (HiCMamba), reducing its generalizability to other genomic enhancement methods.

Provenance

Source

figshare, authored by Minghao Yang.

Collection Method

Generated from the ablation study experiments of the HiCMamba deep learning model.

Freshness

Last updated on March 24, 2026.