. 2023 Jan 9;55(1):26–33. doi: 10.1038/s41588-022-01267-w

Table 1.

Current guidance by various stakeholder and academic groups on the number of specimens to sequence for detection of novel variants at low prevalence

	Recommendation on number/proportion of positive specimens to sequence	Critical considerations
World Health Organization/European Centre for Disease Prevention and Control^7,8	Minimum number of sequences to detect at 1% variant proportion with 95% confidence for given number of reported cases: • 141 (<1,000 cases) • 196 (1,001–2,500 cases) • 243 (2,500–5,000 cases) • 270 (5,001–10,000 cases) • 285 (>10,000 cases)	• Agnostic to variant properties • Assumes specimen pool to be sampled for sequencing is representative of circulating diversity but acknowledges that, unless testing coverage is evenly distributed, this will be a biased sample • Notes that, in countries with limited sequencing capacity, monitoring relative prevalence of variants should be prioritized
Brito et al.⁵	At least 0.5% of all cases, with a turnaround time of 21 days to detect novel lineage before it reaches 100 cases at 20% probability	Based on sequencing data from Denmark, which is testing at an average of >2,000 tests per 100,000 persons per day⁹
Wohl et al.⁶	1–29 sequences per day to detect an Alpha-like variant based on 0.03% initial introduction for a population of 10,000 (assuming growth rate of 0.1 per day) at 1% variant proportion with 95% confidence^a	• Assumes that the observed variant proportion in the positive specimens collected is representative of the circulating variant proportions among the infected population. This requires a large number of specimens that are randomly collected for assumption to hold true at low circulating variant proportions • A correction factor is included to correct for biases in the observed variant proportion, but only pertaining to those arising from the relative differences in diagnostic sensitivity, sample qualities and conditional asymptomatic and symptomatic testing probabilities between the two circulating variants

Recommendation on number/proportion of positive specimens to sequence

Critical considerations

World Health Organization/European Centre for Disease Prevention and Control^7,8

Minimum number of sequences to detect at 1% variant proportion with 95% confidence for given number of reported cases:

• 141 (<1,000 cases)

• 196 (1,001–2,500 cases)

• 243 (2,500–5,000 cases)

• 270 (5,001–10,000 cases)

• 285 (>10,000 cases)

• Agnostic to variant properties

• Assumes specimen pool to be sampled for sequencing is representative of circulating diversity but acknowledges that, unless testing coverage is evenly distributed, this will be a biased sample

• Notes that, in countries with limited sequencing capacity, monitoring relative prevalence of variants should be prioritized

Brito et al.⁵

At least 0.5% of all cases, with a turnaround time of 21 days to detect novel lineage before it reaches 100 cases at 20% probability

Based on sequencing data from Denmark, which is testing at an average of >2,000 tests per 100,000 persons per day⁹

Wohl et al.⁶

1–29 sequences per day to detect an Alpha-like variant based on 0.03% initial introduction for a population of 10,000 (assuming growth rate of 0.1 per day) at 1% variant proportion with 95% confidence^a

• Assumes that the observed variant proportion in the positive specimens collected is representative of the circulating variant proportions among the infected population. This requires a large number of specimens that are randomly collected for assumption to hold true at low circulating variant proportions

• A correction factor is included to correct for biases in the observed variant proportion, but only pertaining to those arising from the relative differences in diagnostic sensitivity, sample qualities and conditional asymptomatic and symptomatic testing probabilities between the two circulating variants

^aWe used the spreadsheet (https://github.com/HopkinsIDD/VOCsamplesize) provided and input appropriate parameters to obtain the recommendation relevant to the simulated epidemics.