doubletrouble

Workflows

Standard Workflow

library(doubletrouble)
library(syntenet)

# Load example data
data(yeast_seq)
data(yeast_annot)
data(diamond_intra)

# Prepare input data
pdata <- process_input(yeast_seq, yeast_annot)

# Classify gene pairs using the standard scheme
c_standard <- classify_gene_pairs(
  annotation = pdata$annotation,
  blast_list = diamond_intra,
  scheme = "standard"
)

# Calculate duplicate frequencies
table(c_standard$Scerevisiae$type)

Note: Inputs are processed protein sequences and gene annotations along with intraspecies DIAMOND search results; output is a list of classified duplicated gene pairs.

When to Use

• To identify and classify duplicated genes from whole-genome protein sequences using classify_gene_pairs.
• To classify duplicates into different schemes: binary (SD vs SSD), standard (SD, TD, PD, DD), or extended (SD, TD, PD, TRD, DD).

When NOT to Use

• For finding synteny or running sequence similarity searches directly; use syntenet functions like process_input and run_diamond instead.
• When whole-genome protein sequences (proteome) and gene annotations (GFF3/GTF) are not available.

Data Requirements

• Proteome: List of AAStringSet objects containing translated sequences of the primary transcripts.
• Annotation: List of GRanges objects containing genomic coordinates of all features.
• Intraspecies BLAST/DIAMOND results: A list of data frames with tabular output (e.g., from run_diamond).

Key Parameters

• annotation: Processed annotation list (a GRangesList object).
• blast_list: A list of data frames with DIAMOND/BLAST tabular output for intraspecies comparisons.
• scheme ("standard"): Classification scheme to use ("binary", "standard", "extended", or "full").
• blast_inter: List of data frames with DIAMOND/BLAST tabular output for interspecies comparisons (required for "extended" and "full" schemes).

Best Practices

• Ensure list names in the annotation match the list names in the sequence object using setequal(names(seqs), names(annotation)).
• Keep only the longest sequence for each protein-coding gene to avoid isoforms.
• Perform bidirectional similarity searches for interspecies comparisons and collapse them using collapse_bidirectional_hits before classification.

Common Pitfalls

• Mismatching list names between sequences and annotations: Ensure names are consistent across lists.
• Including multiple isoforms per gene: Filter the input sequences to keep only the longest sequence per gene.

Alternatives

• syntenet for synteny detection and running DIAMOND alignments.

Citations

• Ohno 2013, Springer Science & Business Media.
• Buchfink, Reuter, and Drost 2021, Nature Methods.

References

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