Scanmir

scanmir

Workflows

Standard Workflow

A set of tools for working with miRNA affinity models (KdModels), efficiently scanning for miRNA binding sites, and predicting target repression. It supports scanning using miRNA seeds, full miRNA sequences (enabling 3' alignment) and KdModels, and includes the prediction of slicing and TDMD sites. Finally, it includes utility and plotting functions (e.g. for the visual representation of miRNA-target alignment).

library(scanMiR)
data("SampleTranscript")
data("SampleKdModel")

# Scan for matches using a KdModel
matches <- findSeedMatches(SampleTranscript, SampleKdModel, verbose = FALSE)

# View target alignment
viewTargetAlignment(matches[1], SampleKdModel, SampleTranscript)

# Aggregate matches
agg_matches <- aggregateMatches(matches)

Input: A transcript sequence (DNAStringSet or character) and a miRNA KdModel; Output: A GRanges object of matches and a data.frame of aggregated repression values.

When to Use

• To scan sequences (character vector or DNAStringSet) for miRNA binding sites using a seed sequence, full miRNA sequence, or a KdModel using findSeedMatches.
• To predict dissociation constants (Kd) and binding types for specific 12-mer sequences using assignKdType.
• To visualize miRNA-target alignments (including 3' supplementary pairing) using viewTargetAlignment or plot KdModel affinities using plotKdModel.
• To aggregate predicted miRNA repression across transcripts using aggregateMatches.

When NOT to Use

• For general RNA-seq differential expression analysis without miRNA target prediction (use packages like DESeq2 or edgeR).
• For predicting miRNA-target interactions without sequence-level binding site information or affinity models.

Data Requirements

• miRNA seeds (character vector of length 7 or 8), full miRNA sequences, or KdModel / KdModelList objects.
• Target transcript sequences as a character vector or a DNAStringSet (optionally with ORF.length metadata column).

Key Parameters

• verbose (TRUE/FALSE): Controls progress reporting during scanning in findSeedMatches.
• onlyCanonical (FALSE): Restricts the scan to canonical miRNA binding sites when using a KdModel.
• ret ("GRanges"): Specifies the return format of findSeedMatches (e.g., "GRanges", "data.frame", or "aggregated").
• shadow (0): Treats matches within the first shadow positions of the UTR as if they were in the ORF.
• minDist (7): Minimum distance between matches of the same miRNA; only the highest affinity match is kept.
• what ("seeds"): Specifies what to plot in plotKdModel (e.g., "seeds").

Best Practices

• Provide the ORF.length as a metadata column in the input DNAStringSet to distinguish between matches in the ORF and 3'UTR regions.
• Use onlyCanonical = TRUE in findSeedMatches if you want to exclude low-affinity non-canonical binding sites.
• Use aggregateMatches to compute the predicted repression of transcripts based on the biochemical model of occupancy.
• For large scans, use multithreading by passing a BiocParallel parameter (e.g., BP = MulticoreParam()) to findSeedMatches.

Common Pitfalls

• Providing the seed sequence in the wrong orientation: Ensure the seed is given as it would appear in the target sequence (reverse complement of the miRNA seed).
• Memory exhaustion during large scans: Limit the number of seeds processed simultaneously using the n_seeds parameter or set useTmpFiles = TRUE.

Alternatives

• targetscan: For standard TargetScan-based miRNA target predictions.
• mirbase: For retrieving miRNA sequences and annotations.

Citations

• McGeary, Lin et al. (2019), Science (for the biochemical model of miRNA target repression).
• Grimson et al. (2007), Molecular Cell (for canonical site types and shadow effect).

References

• Homepage: bioconductor.org/packages/scanmir
• Vignette: https://bioconductor.org/packages/release/bioc/vignettes/scanmir/inst/doc/scanmir.html