Fig 1.
(a), MPCR bias was evaluated using three T cell populations spiked with cells of known TCRs and three controls (with replication) pooled with 33 plasmids inserting known TCR sequences. (b), Comparison of MPCR and 5’RACE. RNA was derived from the PBMC of three healthy donors. It was used to profile the TRB repertoire via both MPCR and 5’RACE. The analyses included a calculation of MPCR/5’RACE reproducibility and comparison of the repertoire characteristics between 5’RACE and MPCR.
Fig 2.
MPCR bias evaluated by the mixed plasmid samples.
Thirty-three plasmids were mixed in the sample with three different pooling gradients (a, b, and c represent plasmid mixes 1, 2, and 3, respectively, details in S3 Table). Their replications are showed in d, e, and f. The observed frequencies of the thirty-three templates were calculated and compared with the expected pooling frequencies. The numbers in the top left corner (b, c, e, and f) represent the Pearson correlation coefficients between the observed and expected frequencies.
Fig 3.
Expected versus observed frequencies for 5 spiked-in clones in three mixtures.
Five CD4+ T cell clones (A, B, C, D, and G) were spiked into sorted CD4+ T cell populations at concentrations that spanned five orders of magnitude. TRB CDR3 was amplified via MPCR, and the observed clone frequency was calculated. The results demonstrated a high concordance between the expected and observed clone frequencies in three samples (Mix1, Mix2, and Mix3).
Table 1.
Statistics of the multiplex PCR and 5’RACE libraries.
Fig 4.
Correlation and consistency in MPCR/5’RACE replicate samples.
(a) The correlation of the CDR3 AA frequencies in two MPCR replicates. Each dot represents a unique CDR3 AA sequence. Shared CDR3 sequences are indicated in blue, and specific sequences are indicated in red. The number in the top left corner is the Pearson correlation coefficient of the two data sets. (b). The consistency of the CDR3 sequences between two MPCR replicates. CDR3 AA sequences are ranked by frequencies in each replicate. We subsequently compared the shared CDR3s between the two replicates beginning with the top ranking clones, e.g., top 10 and top 100. The number (right Y axis) and percentage of shared sequences (left Y axis) are calculated and shown in the solid black dot and blue circle, respectively. (c). The correlation of the CDR3 amino acid sequences in two 5’RACE replicates. (d). The consistency of the CDR3 sequences between two 5’RACE replicates.
Fig 5.
Frequency distribution of TRBV and TRBJ genes for both MPCR and 5’RACE samples.
TRB of three samples were amplified by both MPCR (trilateral) and 5’RACE (square). All V and J frequencies were normalized. The V and J frequencies were compared between the MPCR and 5’RACE samples. Flag: *, p-value < 0.05. **, p-value <0.01, paired t test. AFD, average frequency difference between the 5’RACE and MPCR samples (AFD = f(MPCR)-f(5’RACE)). +, AFD>0; -, AFD<0. (a) Frequencies of 13 TRBJ genes. (b) Frequencies of 48 functional TRBV genes.
Fig 6.
Correlation and consistency between MPCR and 5’RACE samples.
(a, b, c), the consistency of S01, S02, and S03. CDR3 AA sequences are ranked by frequencies in each replicate. We subsequently compared the shared CDR3s between the two replicates beginning with the top ranking clones, e.g., top 10 and top 100. The number (right Y axis) and percentage of shared sequences (left Y axis) are calculated and shown in the solid black dot and blue circle, respectively. (d, e, f), the correlations among S01, S02, and S03. The blue points represent the shared CDR3 in both samples; the red points represent the specific CDR3 in each sample. One extreme point was removed from S02 and S03. The number in the top left corner is the Pearson correlation coefficient of the two data sets.
Fig 7.
Comparison of V-J pairing coverage between MPCR and 5’RACE.
Light gray represents the number of covered V-J pairings for 5’RACE samples; dark gray represents the number for MPCR samples.
Fig 8.
Comparison of PCR/Sequencing errors induced by MPCR and 5’RACE.
The mismatch derived from the alignment results of the arrangement sequence to the V/J germline genes and alleles (IMGT). The left indicates the percentage of mismatched nucleotides in total nucleotides (divided by the total bases); the right indicates the rate of the sequences that contain mismatches in total sequences (divided by total sequences).