V-REGION mutation levels of IGH rearrangements of known isotype
Immunoglobulin heavy chain rearrangements were sequences from PBMCs of 8 individuals using IGHC primers for all isotypes bar IgE. Each indivdual has a unique label 'Subject' identifier of the format NSXXX. Sequencing was carried out on the 454 pyrosequencing platform and rearranged immunoglobulin heavy chain sequences were partitioned against germline IGHV, IGHD and IGHJ using iHMMune-align.
The datasets for each individual has been filtered to remove non-productive rearrangements (frame shift of the IGHJ or stop codons), those with greater than 45 IGHV mutations, sequences which contained indels in the V or J, and those with ambiguities. Within each data only unique sequences were reatained (100% identical) with a 'read copy no.' assigned to track the underlying copies of a sequence.
Clonally-related sequences within each subject's dataset were also identified. Following germline reversion of all non-CDR3 nucleotide positions, sequence were clustered using the vmatch package (http://www.vmatch.de/) allowing up to 3 differences. A single representative for each cluster (representing a likely clonal set) was retained within the dataset. Representative sequences were selected based on the sequence with the highest copy number, or in cases where sequences shared the highest copy number, the least mutated sequence. If a cluster spanned multiple isotypes then a representative for each isotype was kept. Clusters are identified by a label that includes the subject in which the cluster was identified. Sequences that weren't part of a cluster as labelled with the subject identifier and "na". Where a cluster spans multiple isotypes the 'Span multiple isotypes' column is set to TRUE.
Replacement (R) and silent (S) mutation counts for complementarity determining regions (CDRs) and framework regions (FR) were determined for each sequence. CDR definitions that take the outer limits of the Kabat and IMGT systems were used. Where more than one mutation occurred within a single codon, all independent pathways to the final mutational outcome were considered and the R and S counts were weighted accordingly.
For sequences with an Mv (total V-REGION mutation from codon 25 through 104, that is, excluding the CDR3 contribution of the V-REGION) greater than zero the proportion of total mutations that represent R in CDRs (CDR1 and CDR2) was calculated. These values were used in the analysis of antigen selection between sequences of different isotypes.
V-REGION mutation frequencies and CDR3 amino acid sequences and lengths are included.
