Sample collection and transport
Between February 29 and March 28, 2020, samples were obtained from patients with suspected COVID-19 from the Tamale Teaching Hospital (TTH) in the Northern Region, and Komfo Anokye Teaching Hospital (KATH) and Kumasi South Hospital (KSH) in the Ashanti Region of Ghana (Fig. 1). Recruitment of cases at these facilities was done according to the Ghana National Surveillance Strategy protocol . Suspected COVID-19 cases were defined as individuals presenting with fever (>38 °C) or a history of fever and symptoms of respiratory tract illness such as cough or shortness of breath, or individuals who were in close contact with a person who was suspected or confirmed to have COVID-19.
Nasopharyngeal and or oropharyngeal swabs were obtained using flocked swabs (Copan Group, Brescia, Italy) and kept in 500 µl of RNAlater (QIAGEN, Hilden, Germany) in 1.5-ml tubes (Eppendorf, Regensburg, Germany) and transported immediately at ambient temperature for confirmation at the Kumasi Centre for Collaborative Research in Tropical Medicine (KCCR), Kumasi, in the Ashanti Region of Ghana.
The KCCR is one of the two main research laboratories in Ghana designated for COVID-19 testing. This is because of its longstanding experience and the expertise of its scientists in studies related to coronaviruses (http://kccr.org/). The Centre is currently Ghana’s second largest testing site, which serves the northern sector of the country. During the early phase of the pandemic, the Centre received samples from 12 out of 16 regions in Ghana and tested approximately 1,200 samples daily.
Viral RNA extraction and PCR detection
Using a starting volume of 140 µl, both nasopharyngeal and oropharyngeal swabs from each patient were extracted together as a single sample using a QIAGEN Viral RNA Mini Kit (QIAGEN, Hilden, Germany) according to manufacturer’s instructions. Samples were eluted in a 100-µl volume, and SARS-CoV-2 RNA was detected using a RealStar® SARS-CoV-2 RT-PCR Kit (Altona, Germany) according to the manufacturer’s instructions. Sample quantification was done using an externally generated standard curve based on serially diluted SARS-CoV-2 in vitro transcripts. All samples with a cycle threshold (Ct) of 40 or above were considered negative. All amplification runs were validated by including positive and negative controls.
High-throughput sequencing for samples with sufficiently high RNA concentrations as determined by quantitative real-time PCR were sequenced using an Illumina NextSeq platform (Illumina, San Diego, California, U.S.) and a KAPA RNA Hyper Prep kit (Roche Molecular Diagnostics, Basel, Switzerland) according to manufacturer’s instructions. In order to estimate the potential impact that long-distance transport of samples to testing centers may have on sequencing results, the two samples that were closest in terms of viral RNA concentration were tested using an Agilent 4200 TapeStation system (Agilent Technologies, CA, USA). These samples comprised one from TTH in the north, which is approximately 400 kilometers from the testing site, and another from KATH in Kumasi (in the same city as the testing site, approximately 9 km away).
Sequences from this study, with the exception of three that had less than 80% of the genome sequenced, were compared to previous sequences from Ghana and representative sequences from regions where patients had previously travelled. These included sequences from the USA, Japan, France, and Guinea and available sequences from sub-Saharan Africa. All sequences from the region of interest as of April 2020 were obtained from GISAID (https://www.gisaid.org/), and duplicates were removed. The non-redundant sequences were then clustered at a minimum threshold of 99.9% using CD-HIT (http://weizhongli-lab.org/cd-hit/), and representative sequences from each cluster were selected. Multiple sequence alignments were done using the MAFFT plugin in Geneious prime (http://www.geneious.com). Phylogenetic analysis was done by Bayesian inference using the MrBayes  plugin in Geneious prime with a chain length of 1.1 million and a subsampling frequency of 200. A general time-reversible substitution model with a gamma distribution and proportion of invariable sites (GTR+G+I) was used for the analysis. All sequences with genome coverage greater than 80% were analyzed using the Phylogenetic Assignment of Named Global Outbreak Lineages (PANGOLIN) online resource (https://pangolin.cog-uk.io/) for lineage assignment.
We obtained ethical approval from the Committee on Human Research Publications and Ethics of the School of Medicine and Dentistry at the Kwame Nkrumah University of Science and Technology (CHPRE/AP/462/19) and the Institutional Review Board of the Ghana Health Service (GHS-ERC087/03/20).