SUN-031 Quantification of Pooled Libraries for Optimizing Cluster Density in Next Generation Sequencing

Background: Next-generation sequencing (NGS) consists of massive parallel processing libraries where millions of reads are generated in a single sequencing run. To ensure efficient sequencing, the amount of DNA pool libraries must be measured precisely because it influences the density of the clusters formation. There are different quantification methods aiming to obtain high density clusters with Q30>80%, avoiding the negative effects of overclustering. Our aim was to compare the density of the clusters in exome libraries quantified by two different methodologies qPCR and MiSeq Nano V2 method. Exome libraries were prepared using SureSelect Human All Exons kits (Agilent Technology) and sequenced on Illumina HiSeq2500 platform (Illumina) using the V4 high output sequencing kit. The pools were prepared in equimolar concentrations using qPCR (KAPA Library Quantification Kit), to achieve final pool library concentration of 10 nM. Three groups were defined: the original 10 nM group (n=16); a qPCR group (n=16) in which the pool concentrations were corrected by qPCR and the MiSeq Nano V2 group (n=12) where pool concentrations were corrected by clustering data of MiSeq Nano in a direct linear proportion. The statistics analysis was performed by Kruskal-Wallis followed by Dunn test. Cluster density values are reported as the absolute distance from the optimal density (1000 k/mm(2) of flow cell) specified for V4 kit. Results: The 10 nM group showed a median of 191, iqr (interquartile range) = 48-273, the qPCR group the median was 80, iqr = 30-135 and the MiSeq Nano V2 group the median was 46, iqr = 13-78. Therefore, the MiSeq NanoV2 method was significantly better than the other groups at either generating closer distances from the target cluster density value (Kruskal-Wallis test, p<0.05, H=7.967, p=0.02 with Dunn’s test, mean rank difference=13.84, p=0.01). The rate of success in the generation of data without overclustering effects was also better in Miseq Nano V2 group (Kruskal-Wallis test, p<0.05, H=7.525, p=0.02, Dunn’s multiple comparison test, mean rank difference=-9.625, p=0.01. Conclusion: The MiSeq Nano V2 method is significantly more precise to quantify the pool library concentration, increasing data yield per run, reducing potential loss of data due to inadequate clustering and covering, saving time and money in clinical research laboratories.