# Bioconductor Tfbstools

> TFBSTools is a package for the analysis and manipulation of transcription factor binding sites. It includes matrices conversion between Position Frequency Matirx (PFM), Position Weight Matirx (PWM) and Information Content Matrix (ICM). It c

- Skill: `biomate-ai/bioconductor-tfbstools` (Agent Skill)
- Install (CLI): `npx skillmds@latest add biomate-ai/bioconductor-tfbstools`
- Raw SKILL.md: https://api.skillmd.com/api/skills/biomate-ai/bioconductor-tfbstools/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: biomate-ai (https://skillmd.com/u/biomate-ai)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/biomate-ai/bioconductor-tfbstools

---


# TFBSTools

## Dependencies & Environment

> Package-intrinsic requirements from the Bioconductor landing page — reproduce in any R environment.

- **Version:** 1.50.0 · **Bioconductor:** 3.23 · **R:** ≥ 4.6
- **Imports:** Biobase, Biostrings, pwalign, BiocGenerics, BiocParallel, BSgenome, caTools, DirichletMultinomial, Seqinfo, GenomicRanges, gtools, IRanges, DBI, RSQLite, rtracklayer, seqLogo, S4Vectors, TFMPvalue, XML, XVector
- **Install:** `BiocManager::install("TFBSTools")`

## When to Use
- **Matrix Conversion**: Converting Position Frequency Matrices (PFMs) to Position Weight Matrices (PWMs) or Information Content Matrices (ICMs) using `toPWM` and `toICM`.
- **Motif Scanning**: Scanning nucleotide sequences or pairwise alignments for transcription factor binding sites using `searchSeq` or `searchAln`.
- **Database Querying**: Querying and retrieving matrix data from the JASPAR database using `getMatrixSet`.
- **Random Profile Generation**: Generating random profile matrices via permutation or Dirichlet multinomial mixture models using `permuteMatrix` and `rPWMDmm`.
- **Visualization**: Visualizing sequence logos for basic PWMs or Transcription Factor Flexible Models (TFFMs) using `seqLogo`.

## When NOT to Use
- For high-throughput scanning of thousands of large genomic regions, use `motifmatchr` instead, because `searchSeq` is optimized for targeted `DNAString` or `DNAStringSet` objects rather than massive `GRanges` peak sets.
- For de novo motif discovery without external dependencies, use `universalmotif` because `runMEME` requires the external MEME suite to be installed on the system.

## Data Requirements
- **Input Format**: `PFMatrix` objects, `DNAString` or `DNAStringSet` for sequences, and `Axt` objects for alignments.
- **Structure**: 4-row matrices representing A, C, G, T counts or probabilities.
- **Normalization State**: Raw position frequency counts for PFMs, which are then converted to log2 probability ratios for PWMs.

## Key Parameters
- **type** (default): The type of matrix conversion in `toPWM` (e.g., "log2probratio") or the type of p-value calculation in `pvalues` (e.g., "TFMPvalue").
- **pseudocounts** (default): Numeric value (default 0.8) added to correct small counts or eliminate zero values before log transformation in `toPWM`.
- **bg** (default): A named numeric vector representing background nucleotide frequencies (e.g., `c(A=0.25, C=0.25, G=0.25, T=0.25)`).
- **min.score** (default): Minimum score threshold (e.g., "60%" or "80%") for a sequence match to be reported in `searchSeq` or `searchAln`.
- **schneider** (default): Logical in `toICM` indicating whether to apply the Schneider correction.
- **strand** (default): Controls which strand is searched in `searchSeq` (e.g., "*" for both strands).

## Best Practices
- Use `pseudocounts=0.8` in `toPWM` rather than the square root of sequences to avoid overly harsh corrections on zero values.
- When scanning sequences with `searchSeq`, use `pvalues` with `type="TFMPvalue"` to calculate empirical p-values for the match scores.
- Use `writeGFF3` or `writeGFF2` to export `SiteSet` objects into standard genomic formats for downstream visualization.

## Common Pitfalls
- **Pitfall**: `runMEME` fails to execute.  
  *Fix*: Ensure the external MEME software suite is installed and the `binary` argument correctly points to the executable.
- **Pitfall**: `searchSeq` returns too many false positive hits.  
  *Fix*: Increase the `min.score` threshold (e.g., from "60%" to "90%") to enforce stricter matching.
- **Pitfall**: `toPWM` throws an error due to zero counts.  
  *Fix*: Ensure `pseudocounts` is set to a positive value (like 0.8) to eliminate zero values before log transformation.

## Alternatives
- **motifmatchr**: For faster, parallelized motif scanning across large `GRanges` objects.
- **universalmotif**: For comprehensive motif manipulation and de novo discovery without relying strictly on external binaries like MEME.
- **Biostrings**: For basic PWM matching (`matchPWM`), though it lacks the advanced statistical models and TFFM support of `TFBSTools`.

## Citations
- Tan G, Lenhard B (2016). TFBSTools: an R/Bioconductor package for transcription factor binding site analysis. *Bioinformatics*.
- Mathelier and Wasserman (2013). The next generation of transcription factor binding site, TFFM.

## References
- Homepage: https://bioconductor.org/packages/TFBSTools
- Vignette: https://bioconductor.org/packages/release/bioc/vignettes/TFBSTools/inst/doc/TFBSTools.html

<!-- biomate-cta -->
---

## Run this on BioMate

This skill is the **knowledge layer** — when, why, and how to use `tfbstools`. To **run this analysis on your own data** with managed compute, automated QC, and reproducible outputs, use **[BioMate](https://www.biomate.ai?ref=kb&pkg=tfbstools)** — free to start.

▶ **[Open `tfbstools` on BioMate →](https://www.biomate.ai?ref=kb&pkg=tfbstools)**