Fig. 4: CellectSeq motif-based Ab discovery strategy of CD151 selective Abs.

A Scoring Specificity: (i) Consider the candidate antibody NGS clone, the CDR sequences and counts in the positive (POS) and the negative (NEG) selection pools at round 4 are reported. (ii) We enumerate all consensus motifs in the CDR sequences. (iii) We score the binding specificity of the candidate antibody by assessing the separation between the distributions of the frequencies of the motifs in the positive (POS) and negative (NEG) pools. To this end, we calculate the p value of the t-test (Methods). B, C In each figure, the abundance of each sequence in the positive pool (positive Counts, x-axis) is plotted versus the ratio of the abundance in the positive pool over the negative pool (positive/negative, y-axis). Each circle represents one unique Ab clone, colored circles (except black) represent families of homologous sequences (sequence identity >0.75). The dashed red lines define an upper-right quadrant that contains enriched Ab clones defined as occurring more than 200 times in the positive pool and being greater than fourfold enriched relative to the negative pool. In contrast to Fig. 3, the dashed red lines here serve only as visual references to inspect the individual enrichment levels of the clones predicted by CellectSeq as highly specific (p value <10⁻¹⁰). B Before Filtering: NGS sequences and predicted sequences before filtering hybridization errors (N:368,427). (i) Distribution of unique sequences in the positive pool and their enrichment over the negative pool. (ii) Distribution of predicted sequences with high specificity (p value <10⁻¹⁰), colored by family of homologs. (iii) Distribution of selected sequences for in situ validation, colored by family of homologs. C After filtering: NGS sequences and predicted sequences after filtering hybridization errors (N:105,434). (i) Distribution of unique sequences in the positive pool and their enrichment over the negative pool. (ii) Distribution of predicted sequences with high specificity (p value <10⁻¹⁰), colored by family of homologs. (iii) Distribution of selected sequences for in situ validation, colored by family of homologs. Colored circles represent Ab clones named CD151-1 to CD151-4 which were selected and validated as specific CD151 binding Abs by cellular phage ELISA. Clones CD151-2 to CD151-4 are in the top-left quadrant, which means they are outside the scope of the NGS enrichment ranking method. D Before versus After filtering: Difference between number of unique sequences and prediction results before and after filtering hybridization errors. (i) Number of unique sequences in the positive pool before and after filtering. (ii) Number of predicted sequences before and after filtering. (iii) Number of validated sequences before and after the filtering.