How JOINSOLVER^® Works

Introduction

JOINSOLVER^® was developed specifically to analyze the CDR3 region of the immunoglobulin genes in human B cells. The strategy of JOINSOLVER^® is to search for D germline sequences flanking V_H and J_H germline genes. Additionally, it searches for P and N type additions in the V_HD and DJ_H junctions. The human D germline gene database employed includes all D segments from the IMGT databank as well as the reverse and DIR germline genes.

Assignment of the 5' boundary of the CDR3_H

JOINSOLVER^® initially interrogates the sequence to find the beginning of the CDR3_H region as codon 93. This codon was used to define the beginning of the CDR3 based on the results of structural analyses of V_HDJ_H rearrangements (2, 3) as recommended (3-5). To identify this codon, JOINSOLVER^® searches for the sequence, “TAT TAC TGT”, which comprises codons 90 to 92 of the V_H region (after Kabat et al (1) and is a conserved motif in most of the human V_H germline genes. If a “TAT TAC TGT” motif is not found, the search is reinitiated with one base-pair change allowed in the sequence. If a “TAT TAC TGT” with one nucleotide change is not found, then homologies with the germline genes are used to find the most likely start of the CDR3_H region. If the start of the CDR3_H region is not yet identified, JOINSOLVER^® marks the CDR3_H as not found and defers finding the CDR3_H region until after V and J matching.

Assignment of the 3' boundary of the CDR3_H

After the V_H end of the CDR3_H is defined, JOINSOLVER^® screens for the J_H border of the CDR3_H. A “C TGG GG” motif demarks the 3’ end of the CDR3_H region and is conserved in all J_H sequences. A similar algorithm is used to find the “C TGG GG” at the 3’ end of the CDR3_H.

Assignment of the V_H, D and J_H segments

Once the CDR3_H region is identified, V_H, J_H and D assignment is carried out. The V region is matched to a database of germline genes from the “TAT TAC TGT” back 3’→5’ toward the beginning of the sequence, and forward in the 5’→3’ direction to the end of the germline gene. The J_H region is matched from the “C TGG GG” back to the beginning of the germline gene and forward until the end of the sequence or the end of the germline gene is identified. The V_H and J_H regions are scored with an alignment score that assigns a +5 to a nucleotide match and -4 for a mismatch between the unknown sequence and the germline (7).

The end of the V_H region is identified when the given unknown sequence matches the complete V_H germline gene or has a mismatch after the “TAT TAC TGT” with the highest scoring V_H germline. The beginning of the J_H region is defined when the unknown sequence has 1 mismatch before the “C TGG GG” with the highest scoring J_H region or the sequence matches the complete J H germline gene.

In the event that the CDR3_H was inititally not found, JOINSOLVER^® looks for matches between the V and J germline databases and the unknown sequence. The unknown sequence is aligned to the highest scoring germline genes. The CDR3_H region is defined as the region from codon 93 and the “C” of the “C TGG GG” motif. The V_H end and J_H start are defined the same way as if the CDR3_H region had been found first.

After V_H and J_H assignment, D segment assignment is carried out using a consecutive match scoring system. All matches to the D germline genes are scored and sorted based on the V_H-J_H distance (the distance in nucleotides between the end of the V_H segment and the beginning of the J_H segment). The longest matches are aligned and returned to the user.

A Monte Carlo simulation was used to determine the minimal length of a D segment match necessary to ensure that the match was unlikely to be a random occurrence. As shown in the table below, the match length required for identification depends on the V_H-J_H distance. Seven to 11 consecutively matching base pairs are necessary to identify a D segment with sufficiently high probability that it is unlikely to be from random chance.