Table 1.

Characteristics of the datasets included in meta-analysis of candidate whole blood transcriptional signatures for incipient tuberculosis

	Samples included	Study design	Population	Setting	HIV status	Sampling	Follow-up duration and method	Tuberculosis case definition	RNA sequencing methods	Newcastle-Ottawa Scale score	Baseline tuberculosis assessment
London tuberculosis contacts26	324 (8 tuberculosis; 316 healthy)	Cohort	Adult tuberculosis contacts	London, UK	Negative	Baseline	Median 1·9 (IQR 1·7– 2·2) years, record linkage	Culture-confirmed, or clinically diagnosed	15–20 million 41 bp paired-end reads	7/7	Clinical evaluation and chest x-ray
Adolescent Cohort Study24	287 (73 tuberculosis; 214 healthy)	Nested case-control	Adolescents with latent tuberculosis infection	South Africa	Negative	Serial (0, 6, 12, and 24 months)	2 years, active	Intrathoracic disease with 2 positive smears, or 1 positive culture	30 million 50 bp paired-end reads	9/9	Clinical evaluation; tuberculosis <6 months from enrolment excluded; chest x-ray not specified
Grand Challenges 6-7425	412 (98 tuberculosis; 314 healthy)	Nested case-control	Adult household pulmonary tuberculosis contacts	South Africa, The Gambia, Ethiopia	Negative	Serial (0, 6, and 18 months)	2 years, active	Culture-confirmed or clinically diagnosed	60 million 50 bp paired-end reads	9/9	Clinical evaluation; tuberculosis <3 months from enrolment excluded; chest x-ray not specified
Leicester tuberculosis contacts17	103 (4 tuberculosis; 99 healthy)	Cohort	Adult tuberculosis contacts	Leicester, UK	Negative	Baseline plus serial for a subset*	2 years, active	Confirmed by culture or Xpert MTB/RIF	25 million 75 bp paired-end reads	7/7	Clinical evaluation and chest x-ray

^*Owing to the high frequency of serial sampling (<6-monthly), only baseline samples were included.