2.4. Technical Reproducibility

For reliable detection of clonotypes activated during infection, pathological processes, or treatment conditions, it is critical to measure the accuracy of AIR repertoire profiling. To achieve this goal, we recommend running each sample (or at least a few samples) if possible, in technical replicates to estimate reproducibility in the detection of high- and medium-abundant clonotypes. For samples with a limited amount of starting RNA, you can split and amplify in replicates the same cDNA sample after the initial forward primer extension step [18].

In addition to technical or amplification replicates, it is desirable to repeat AIR RNA profiling from at least two to three RNA samples independently purified from the same biological sample (e.g., different tissue sections of tumor samples) whenever possible. AIR profiling without replicates may only identify high-abundant clonotypes with a significant loss(false-negative) of medium-abundant clonotypes.
To provide a more quantitative measurement of clonotype abundance level and compensate for variability in the amplification steps (PCR duplicates), both AIR RNA-based and AIR DNA-based assays incorporate UMIs in reverse gene-specific primers. Using UMIs allows one to calibrate and normalize the AIR profiling data (read number of each clonotype) to the copy number of CDR3-specific target mRNA or DNA molecules in the biological sample. Furthermore, to use UMIs for reliable detection and quantitation of high- to medium-abundant clonotypes in the AIR RNA assay, it is important to achieve at least 20-40 reads per UMI, which corresponds to approximately 5-10 million reads/ sample.

In a typical AIR profiling experiment using PBMC or whole blood RNA samples, as illustrated in Fig.13, you should expect to detect the following clonotype classes:

• About 100-500 highly abundant clonotypes (approximately 20% of reads) are amplified from at least 10 mRNA molecules (>10 UMI copies). They can be reproducibly profiled (Fig. 13) from even a single RNA sample without biological triplicates.

• About 3,000-5,000 medium-abundant clonotypes are amplified from at least three mRNA molecules. They could have some significant variation (including false negatives) in measured abundance level due to the low number of CDR3-specific mRNA molecules used in the assay. To reliably measure medium-abundant clonotypes, technical replicates are necessary.

• Approximately 50,000-500,000 clonotypes are present in low abundance, randomly amplified from a large pool of low abundant (1-2 copies) mRNA TCR/BCR molecules (in PBMC or blood samples), and cannot be reproducibly profiled. Comparison of technical AIR repertoire replicates for low abundant clonotypes usually reveals thousands of clonotypes with a minimum overlap in triplicate samples.

Fig 13. Analysis of reproducibility of high-, medium-, and low-abundant clonotypes in whole blood samples measured in triplicates by AIR RNA-based assay in 50 ng of whole blood RNA (10x10^6 reads per sample). Y axis shows the number of CDR3-specific mRNA molecules (based on UMI measurement) amplified for each clonotype with number of unique clonotypes in each class shown on the X axis.