Of the LAMP-sequencing reads obtained, 98% mapped either to the part of the viral genome targeted by the RT-LAMP primers (80.6%) or contained short k-mers derived from primer sequences (17.4%) (fig. S4). This indicated that LAMP-sequencing amplified the targeted sequences. Reads containing only primer sequences were likely to be the result of spurious amplification products as these were also formed in the absence of input RNA (Fig. 1). For quantification of individual LAMP reactions, we classified reads according to whether or not they contained viral sequences, which were not directly covered by the primers (orange segments in fig. S4A), and counted the reads for each sample (as specified by its barcode combination) (fig. S4B). For 754 of the 768 samples, we obtained enough reads to make a call (fig. S5). For the 754 samples that underwent successful LAMP-sequencing, the results confirmed all samples that scored positive on the RT-LAMP assay with a CT < 30 (Fig. 4B and Table 2). For the two samples with a negative RT-qPCR result that scored positive on the RT-LAMP assay (Fig. 3), the LAMP-sequencing call agreed with the RT-qPCR result and thus corrected the RT-LAMP result.