Reproducible codes used for conducting analysis and generating figures and tables of the manuscript Normalization and de-noising of single-cell Hi-C data with BandNorm and scVI-3D. BioRxiv (2021). Ye Zheng*, Siqi Shen* and Sündüz Keleş. * contribute equally.
The software source codes are deposited at
-
BandNorm (R package): https://github.com/sshen82/BandNorm
-
scVI-3D (Python tool): https://github.com/yezhengSTAT/scVI-3D
The curated scHi-C data used in this manuscript are cleaned and organized at https://pages.stat.wisc.edu/~sshen82/bandnorm/ for easy access and direct usage in BandNorm and scVI-3D. This online data repository contains five scHi-C data sets, all at 1Mb. The dataset name, raw data source and reference are summarized below:
-
Ramani2017: GEO accession number GSE84920. Ramani, Vijay, et al. "Massively multiplex single-cell Hi-C." Nature methods 14.3 (2017): 263-266.
-
Lee2019: GEO accession number GSE130711. Lee, Dong-Sung, et al. "Simultaneous profiling of 3D genome structure and DNA methylation in single human cells." Nature methods 16.10 (2019): 999-1006.
-
Li2019: GEO accession number GSE119171. Li, Guoqiang, et al. "Joint profiling of DNA methylation and chromatin architecture in single cells." Nature methods 16.10 (2019): 991-993.
-
Kim2020: https://noble.gs.washington.edu/proj/schic-topic-model/. Kim, Hyeon-Jin, et al. "Capturing cell type-specific chromatin compartment patterns by applying topic modeling to single-cell Hi-C data." PLoS computational biology 16.9 (2020): e1008173.
-
Tan2021: GEO accession number GSE162511. Tan, Longzhi, et al. "Changes in genome architecture and transcriptional dynamics progress independently of sensory experience during post-natal brain development." Cell 184.3 (2021): 741-758.
The metadata summaries that contains the cell barcode, cell type, sequencing depth and sparsity for each data set are also provided at https://pages.stat.wisc.edu/~sshen82/bandnorm/Summary/.
In this repository, the scripts are classified into three subfolders based on the purpose.
-
RunAlgorithms folder contains scripts to run BandNorm, scVI-3D, scHi-C normalization baseline methods that we created for comparison and other publically available tools including scHiCluster, scHiC Topics and Higashi. The outputs of those scripts provide lower-dimensional embeddings and normalized scHi-C counts, which will be used for the downstream analysis and generating figures.
-
AnalysisCodes folder contains analysis related scripts to implement evaluation, comparison and biological discoveries.
-
FigureGeneration folder contains scripts that generate the figures in the manuscript.
For questions, please email Siqi Shen (sshen82@wisc.edu) or submit new issue at this repository.
This BandNorm_and_scVI-3D_manuscript repository is protected by GNU GPL 3 license. Please refer to LICENSE.md for using codes in this repository for your own purposes.