"We think that the most likely scenario is that the SARS-CoV-2 sequence fragments are not from the Antarctic soil samples,but are mistakenly assigned reads from another sample(s) sequenced on the same flow cell at the Sangon Biotech sequencing facility in Shanghai.
The equipment, Illumina HiSeq 4000 that was used for sequencing according to the project’s metadata, is known to be prone to barcode misassignment errors6,7.In a nutshell, as detailed in Illumina’s whitepaper6, this artifact is related to the exponentially increasing density of next-generation sequencing (NGS) chips. Nowadays a single flow cell has such a large capacity that it is rarely filled up by a single sample. For example the 12 antarctic soil samples in PRJNA692319 altogether contain around 320 million paired end reads, while Illumina HiSeq 4000’s capacity is up to 10 billion paired end reads. The increased capacity is utilized through multiplexing, which adds unique sequences, called indexes or barcodes, to each DNA fragment during library preparation.Before final data analysis, sequencing reads from pooled libraries need to be identified and sorted computationally in a process called demultiplexing. This allows large numbers of libraries to be pooled and sequenced simultaneously during a single sequencing run. After adapters are ligated to nucleic acid fragments, the products are cleaned up to remove any free, unligated adapters. Libraries can be cleaned up by a bead-based or gel purification step to remove free adapters or primers. Failure to remove free adapters or primers can lead to contamination of prepared libraries and may result in index hopping and misassignment. As a results, sequencing reads from other, not related samples that are sequenced on the same flow cell can get into our samples."