This is part 2 of 2 of the MRD blog post. (Click here for part 1). In this section, we will discuss how to overcome some of the most common challenges of MRD testing. Overcoming the Challenges In order to mitigate sequencing errors, methods using Molecular Barcodes (MBCs), Unique Molecular Identifiers (UMIs), etc. (which are essentially all the same) may be used, where each starting molecule is sequenced many times. The MBCs are then used to generate consensus sequences from sequences that were likely obtained from the same starting molecule. The assumption is that errors appear due to somewhat stochastic processes and that the consensus sequences will likely be correct. This requires many observations of the same starting molecule, so it will be recommended to generate 10-fold more sequences than there are molecules. Therefore, with 8,000 genomic equivalents, we might want to target a sequencing depth of 80,000. This is a reason why using 10-fold more input ccfDNA may not necessarily be a good thing (in addition to having to obtain a 10-fold larger liquid biopsy) since we may have to increase sequencing depth accordingly to 800,000, which could increase the cost of sequencing 10-fold, which could reduce the likelihood for payment and running the assay profitably.