# Rblast

> rblast

- Skill: `biomate-ai/rblast` (Agent Skill)
- Install (CLI): `npx skillmds@latest add biomate-ai/rblast`
- Raw SKILL.md: https://api.skillmd.com/api/skills/biomate-ai/rblast/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/rblast

---


# rblast

## Workflows

### Create Custom Blast Db

Create a custom BLAST database from user-provided sequences and query it locally.

```r
library(rBLAST)
library(Biostrings)
# Load example sequences
seq <- readRNAStringSet(system.file("examples/RNA_example.fasta", package = "rBLAST"))
# Write to FASTA
writeXStringSet(seq, filepath = "seqs.fasta")
# Create database
makeblastdb("seqs.fasta", db_name = "db/small", dbtype = "nucl")
```
Note: Input is an `XStringSet` object, and output is a set of local BLAST database files written to disk.

### Standard Workflow

Download, extract, and query an existing NCBI BLAST database using local BLAST.

```r
library(rBLAST)
# Download and extract 16S database
tgz_file <- blast_db_get("16S_ribosomal_RNA.tar.gz")
untar(tgz_file, exdir = "16S_rRNA_DB")

# Load existing database
bl <- blast(db = "./16S_rRNA_DB/16S_ribosomal_RNA")

# Query database
seq <- readRNAStringSet(system.file("examples/RNA_example.fasta", package = "rBLAST"))[1]
cl <- predict(bl, seq)
```
Note: Input is a path to a local BLAST database and a query `XStringSet` object; output is a `data.frame` of matches.

### Create And Query Custom Db

Seamlessly interfaces the Basic Local Alignment Search Tool (BLAST) to search genetic sequence data

```r
library(rBLAST)
library(Biostrings)
# 1. Read reference sequences
seq <- readRNAStringSet(system.file("examples/RNA_example.fasta", package = "rBLAST"))
# 2. Write sequences to a FASTA file
writeXStringSet(seq, filepath = "seqs.fasta")
# 3. Create custom BLAST database
makeblastdb("seqs.fasta", db_name = "db/small", dbtype = "nucl")
# 4. Load custom database
db <- blast("db/small")
# 5. Extract a subsequence fragment
fragment <- subseq(seq[1], start = 101, end = 200)
# 6. Query custom database
res <- predict(db, fragment)
```
Note: Input is a set of reference sequences and a query fragment; output is a `data.frame` of alignment matches.

## When to Use
- **Local BLAST Queries**: Running local BLAST queries directly from R using Bioconductor `XStringSet` objects (e.g., `RNAStringSet`, `DNAStringSet`).
- **Custom BLAST Databases**: Creating custom BLAST databases from local FASTA files using `makeblastdb` and querying them locally.
- **NCBI Database Downloads**: Downloading and caching pre-trained NCBI BLAST databases (e.g., 16S rRNA) using `blast_db_get`.

## When NOT to Use
- **Remote NCBI Queries**: For running BLAST queries remotely on NCBI servers without local BLAST+ installation (use `annotate::blastSequences` instead).
- **No Local BLAST+ Installation**: If the BLAST+ command-line tools are not installed on the system or not in the system `PATH`.

## Data Requirements
- Query sequences as `Biostrings::XStringSet` objects (e.g., `RNAStringSet`).
- Reference sequences in FASTA format for database creation.

## Key Parameters
- **db**: Path to the BLAST database in `blast()`.
- **db_name**: Output path and name prefix for the database in `makeblastdb()`.
- **dbtype** ("nucl"): Type of database to create ("nucl" or "prot") in `makeblastdb()`.
- **BLAST_args**: Additional command-line arguments passed to the BLAST executable in `predict()` (e.g., `"-perc_identity 99"`).
- **custom_format**: Custom output format string specifying columns to return in `predict()`.

## Best Practices
- Verify BLAST+ is installed and accessible in R using `has_blast()` or `Sys.which("blastn")`.
- Set the system `PATH` environment variable using `Sys.setenv` if R cannot find the BLAST+ executables.
- Organize custom database files by specifying a directory prefix (e.g., `db/small`) in `makeblastdb`.
- Use `unlink` with `recursive = TRUE` to clean up temporary database directories and FASTA files.

## Common Pitfalls
- **BLAST executable not found**: `Sys.which("blastn")` returns `""`. Fix: Set the `PATH` environment variable to include the BLAST+ installation directory.
- **Database files missing or corrupted**: Occurs if the database path is incorrect. Fix: Ensure the database prefix matches the one specified during `makeblastdb`.

## Alternatives
- **annotate**: Use `blastSequences` for querying the remote NCBI BLAST server.

## Citations
- Altschul, S. F., Gish, W., Miller, W., Myers, E. W., & Lipman, D. J. (1990). Basic local alignment search tool. Journal of Molecular Biology.

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