本文へ移動
cccskills
無料GitHub で公開

bioconductor-hicexperiment

R generic interface to Hi-C contact matrices in `.(m)cool`, `.hic` or HiC-Pro derived formats, as well as other Hi-C processed file formats. Contact matrices can be partially parsed using a random access method, allowing a memory-efficient representation of Hi-C data in R. The `HiCExperiment` class stores the Hi-C contacts parsed from local contact matrix files. `HiCExperiment` instances can be further investigated in R using the `HiContacts` analysis package.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md4.0 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

HiCExperiment

Workflows

Standard Workflow

library(HiCExperiment)
library(HiContactsData)

# Retrieve a sample cool file path
cool_file <- HiContactsData('yeast_wt', format = 'cool')

# Import the contact matrix as a HiCExperiment object
hic <- import(cool_file, format = 'cool', focus = 'I:20001-80000')

# Inspect the imported object
resolution(hic)

Input: A path to a .cool, .mcool, .hic, or HiC-Pro contact matrix file. Output: A parsed HiCExperiment object containing contact frequencies, genomic bins, and metadata.

When to Use

  • Importing Hi-C contact matrices stored in .cool, .mcool, .hic, or HiC-Pro formats using import().
  • Performing memory-efficient random access or subsetting of genomic loci during import using the focus argument.
  • Coercing Hi-C experiment data into standard Bioconductor structures like GInteractions, ContactMatrix, or base matrix using as().

When NOT to Use

  • For downstream Hi-C analysis tasks like compartment calling or insulation score calculation, use HiContacts instead of HiCExperiment alone.
  • For subsetting HiC-Pro formatted matrices during import, as HiCExperiment does not support the focus argument for HiC-Pro files (they must be fully imported into memory).

Data Requirements

  • Input files must be in .cool, .mcool, .hic, or HiC-Pro format (which consists of a .matrix and a .bed file).
  • Optional .pairs or .pairs.gz files can be linked using CoolFile(), HicFile(), or HicproFile() via the pairsFile argument.

Key Parameters

  • format (NULL): Specifies the file format to import (e.g., 'cool', 'pairs').
  • focus (NULL): A character string (e.g., 'I:20001-80000' or 'II:1-500000|II:100001-300000') to subset contacts within a genomic locus of interest.
  • resolution (NULL): Numeric value specifying the resolution at which count values are recovered from multi-resolution files.
  • pairsFile (NULL): Path to a pairs file containing chimeric pairs to associate with the contact matrix.

Best Practices

  • Query available resolutions in multi-resolution files using availableResolutions() before importing.
  • Check available chromosomes in a file using availableChromosomes() to construct valid focus queries.
  • Retrieve the genomic binning structure at the active resolution using bins() or unique regions using regions().

Common Pitfalls

  • Attempting to query subsets of HiC-Pro formatted matrices using focus: This is unsupported and will cause the entire matrix to be imported into memory.
  • Accessing missing scores (e.g., normalized/balanced scores) when they are not present in the input file: Verify available scores using scores(hic) first.

Alternatives

  • InteractionSet: For general representation of genomic interactions without specialized Hi-C file parsers.

Citations

  • Lun, Perry & Ing-Simmons, F1000 Research 2016 (for base Bioconductor classes GInteractions and ContactMatrix).

References

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

The package contains functions for exploratory oligonucleotide array analysis. The dependence on tkWidgets only concerns few convenience functions. 'affy' is fully functional without it.

日本語の概要は準備中です。原文の説明を表示しています。

bioMate-AI/biomate-bioconductor-kb8042026年6月21日 更新

This tool uses the label-tree function from HyPhy to annotate a phylogenetic tree. It allows users to select a subset of leaves using either a regular expression or a list of sequence names, and then apply a specified label to these selecte

日本語の概要は準備中です。原文の説明を表示しています。

bioMate-AI/biomate-bioconductor-kb8042026年6月21日 更新

This package provides a client for the Bioconductor AnnotationHub web resource. The AnnotationHub web resource provides a central location where genomic files (e.g., VCF, bed, wig) and other resources from standard locations (e.g., UCSC, En

日本語の概要は準備中です。原文の説明を表示しています。

bioMate-AI/biomate-bioconductor-kb8042026年6月21日 更新

apeglm provides Bayesian shrinkage estimators for effect sizes for a variety of GLM models, using approximation of the posterior for individual coefficients.

日本語の概要は準備中です。原文の説明を表示しています。

bioMate-AI/biomate-bioconductor-kb8042026年6月21日 更新

ASURAT is a software for single-cell data analysis. Using ASURAT, one can simultaneously perform unsupervised clustering and biological interpretation in terms of cell type, disease, biological process, and signaling pathway activity. Inputting a single-cell RNA-seq data and knowledge-based databases, such as Cell Ontology, Gene Ontology, KEGG, etc., ASURAT transforms gene expression tables into original multivariate tables, termed sign-by-sample matrices (SSMs).

日本語の概要は準備中です。原文の説明を表示しています。

bioMate-AI/biomate-bioconductor-kb8042026年6月21日 更新

AUCell allows to identify cells with active gene sets (e.g. signatures, gene modules...) in single-cell RNA-seq data. AUCell uses the "Area Under the Curve" (AUC) to calculate whether a critical subset of the input gene set is enriched with

日本語の概要は準備中です。原文の説明を表示しています。

bioMate-AI/biomate-bioconductor-kb8042026年6月21日 更新

bioMate-AI のスキルをすべて見る

このスキルの問題を報告する