Allele clusters (ASC / OGRDB)

Some IGHV alleles are nearly identical and are shared between duplicated genes (e.g. IGHV1-69 / IGHV1-69D, IGHV3-23 / IGHV3-23D). These cannot always be assigned unambiguously from short reads, which weakens genotype and haplotype inference. Allele Similarity Clusters (ASCs), from the PIgLET package, group such alleles into clusters with a population-informed threshold and give them a consistent naming scheme (IGHVF<family>-G<cluster>) that is compatible with OGRDB reference sets.

RAbHIT integrates this through a single helper, convertToASC(), which prepares ASC-named data and germline for the standard haplotype workflow.

Optional dependency

This feature requires the piglet package (and a sequence-distance backend such as DECIPHER). Install it with install.packages("piglet"). RAbHIT only loads it when you call convertToASC().

Basic usage

library(rabhit)
data(samples_db, HVGERM)

clip_db <- samples_db[samples_db$subject == "I5", ]

# Convert V calls + germline to ASC names (clusters inferred from the germline)
asc <- convertToASC(clip_db, HVGERM, chain = "IGH")

# Haplotype the ASC-named V genes, anchored on IGHJ6
hap <- createFullHaplotype(asc$clip_db, toHap_col = "v_call",
                           hapBy_col = "j_call", hapBy = "IGHJ6",
                           toHap_GERM = asc$germline, chain = "IGH")

# Plot with the ASC gene order returned by convertToASC()
plotHaplotype(hap, genes_order = asc$genes_order)

convertToASC() returns a list with the ASC-named clip_db, the ASC germline (pass it as toHap_GERM), a genes_order for the plotting functions, and the allele_cluster_table that was used.

Reproducible / OGRDB clusters

By default the cluster table is inferred from the germline, which can vary with the reference set. For reproducible results — and to align with the community reference — supply a fixed ASC table, e.g. the published thresholds from OGRDB/Zenodo:

# Download the published ASC archive and extract its threshold table
archive <- piglet::recentAlleleClusters(get_file = TRUE)
asc_table <- piglet::extractASCTable(archive)

asc <- convertToASC(clip_db, HVGERM, chain = "IGH",
                    allele_cluster_table = asc_table)

You can also pass a table produced once with piglet::inferAlleleClusters() and reuse it across subjects.

Notes

• Clustering is applied to V genes only; D and J calls are left unchanged, so D/J haplotyping continues to use standard IMGT names.
• Duplicated genes are largely resolved for free: alleles shared across duplicated genes fall into the same cluster, removing the cross-assignment ambiguity that RAbHIT would otherwise flag as non-reliable.