Fig. 1: Overall approach and unbiased repertoire sequencing.
a B cells are purified from peripheral blood and mRNA is isolated. Unique molecular identifiers (UMI) are added by template switching reverse transcription (TS-RT) and variable regions are amplified by PCR. Amplicons are sequenced by 2x300 bp Illumina sequencing and UMIs are exploited for error correction. High-quality sequences are used to learn probabilistic models and predict probabilities of bNAb sequences from the CATNAP database. Correlation of probabilities and neutralization allows for identifying highly probable and potent bNAbs. b Biological replicates of 100,000 IgG+ B cells were isolated from three uninfected donors for pipeline validation. A2 represents a technical replicate of A (i.e., the same PCR product, but independent library preparation and sequencing). c Overlap of unique CDRH3s between replicates from samples in b. Upper panel shows the total number of overlapping CDRH3s between biological and technical replicates. Lower panel shows CDRH3 overlap within and between donors as the mean overlap of n = 3623 randomly drawn CDRH3s from each dataset after 100 iterations (see methods for details). d CDRH3 length distributions from samples in b. Upper panel shows overlayed distributions for biological replicates. Lower panel shows p-values from Dunn post-hoc test after global Kruskal–Wallis test. e VH gene nucleotide (nt) mutation frequency distributions from samples in b. Upper and lower panels show overlayed distributions and p-values determined as in d. f Cohort sex and age distributions of n = 57 uninfected individuals for IgG+ repertoire sequencing. g Total and mean number of raw reads, annotated reads, and identical sequences (i.e., the same UMI) that were found at least three times in n = 57 uninfected individuals. h V gene segment distributions, mean CDR3 lengths, and mean V gene mutation frequencies for heavy, kappa, and lambda chains of n = 57 uninfected individuals. V gene segment distributions are depicted as mean values ± SD. Box-plots in f and h depict the 25% and 75% percentiles with the median as average lines and minimum/maximum values as whiskers. CDR complementarity determining region, SHM somatic hypermutation, Rep-Seq repertoire sequencing, HC heavy chain, KC kappa chain, LC lambda chain. Source data are provided as a Source Data file.
