Figure 5: Starting database effect on final output and iteration requirement.
Venn diagrams comparing the number and identity of alleles output by IgDiscover when the following series of starting databases or conditions are utilized: (a) Public RM VH database constructed from previously published rhesus germline V sequences19,43,44,45 versus ‘Single gene RM VH’ database, a VH database containing a single sequence, (VH7.21). (b) Public RM VH database versus ‘Mutated Public RM VH database’, a database created by randomly mutating 10 bases in every sequence within the Public RM VH database. (c) Public RM VH database versus the Human database, VH database created using 351 sequences from the IMGT collection of human VH sequences, none of which are identical to rhesus macaque VH sequences. (d) Public RM VH database versus the Mouse database, VH database created using 407 sequences from the IMGT collection of mouse VH sequences, none of which are identical to rhesus macaque VH sequences. (e) Public RM VH database versus duplicated analysis of Public RM VH database. (f) Public RM VH database using standard J gene settings versus the Public RM VH database when J gene identity is not part of the filtering criteria. (g) Comparison of germline VH alleles identified from F124 IgM library compared with germline VH alleles identified from a combined IgG–IgM library data set. (h) Illustration of the number of iterations to produce a final database from mouse (M1M, M2M, M3M), human (H1M) and rhesus (F132). Each dot indicates the size of the individualized database if the process had been stopped at that iteration and the germline filter applied.
