{"paper_id":"3b147834-112b-4a85-af1b-240a32058196","body_text":"Page 1 of 18\n1 Current sampling and sequencing biases of Lassa mammarenavirus limit \n2 inference from phylogeography and molecular epidemiology in Lassa Fever \n3 endemic regions.\n4\n5 Authors\n6 Liã Bárbara Arruda1b#, Hayley Beth Free2a§, David Simons2§, Rashid Ansumana3, Linzy Elton1, \n7 Najmul Haider 2c, Isobella Honeyborne 1, Danny Asogun 4, Timothy D McHugh 1, Francine \n8 Ntoumi5,6, Alimuddin Zumla1,7, Richard Kock2\n9\n10 Affiliations\n11 1 Centre for Clinical Microbiology, Division of Infection and Immunity, University College \n12 London, London, UK\n13 2 The Royal Veterinary College, University of London, Hatfield, UK.\n14 3 School of Community Health Sciences, Njala University, Bo, Sierra Leone\n15 4 Ekpoma and Irrua Specialist Teaching Hospital, Ambrose Alli University, Irrua, Nigeria.\n16 5 Fondation Congolaise pour la Recherche Médicale (FCRM), Brazzaville, Republic of Congo \n17 6 Institute for Tropical Medicine, University of Tübingen, Germany\n18 7 NIHR Biomedical Research Centre, UCL Hospitals NHS Foundation Trust, London, UK\n19 a Current affiliation Oxford Brookes University, Oxford, UK\n20 b Current affiliation Wellcome Connecting Science, Hinxton, UK\n21 c Current affiliation School of Life Sciences, Faculty of Natural Sciences, Keele University, \n22 Staffordshire, United Kingdom\n23 § Both authors contributed equality to this work\n24 # Corresponding author\n25\n26\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \nNOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.\n\nPage 2 of 18\n27 Abstract\n28 Lassa fever (LF) is a potentially lethal viral haemorrhagic infection of humans caused by Lassa \n29 mammarenavirus (LASV). It is an important endemic zoonotic disease in West Africa with \n30 growing evidence for increasing frequency and sizes of outbreaks. Phylogeographic and \n31 molecular epidemiology methods have projected expansion of the Lassa fever endemic zone \n32 in the context of future global change. The Natal multimammate mouse (Mastomys natalensis) \n33 is the predominant LASV reservoir, with few studies investigating the role of other animal \n34 species. To explore host sequencing biases, all LASV nucleotide sequences and associated \n35 metadata available on GenBank (n = 2,298) were retrieved. Most data originated from Nigeria \n36 (54%), Guinea (20%) and Sierra Leone (14%). Data from non-human hosts (n = 703) were \n37 limited and only 69 sequences encompassed complete genes. We found a strong positive \n38 correlation between the number of confirmed human cases and sequences at the country level \n39 (r = 0.93 (95% Confidence Interval = 0.71 - 0.98), p < 0.001) but no correlation exists between \n40 confirmed cases and the number of available rodent sequences (r = -0.019 (95% C.I. -0.71 - \n41 0.69), p = 0.96).  Spatial modelling of sequencing effort highlighted current biases in locations \n42 of available sequences, with increased effort observed in Southern Guinea and Southern \n43 Nigeria. Phylogenetic analyses showed geographic clustering of LASV lineages, suggestive \n44 of isolated events of human-to-rodent transmission and the emergence of currently circulating \n45 strains of LASV from the year 1498 in Nigeria. Overall, the current study highlights significant \n46 geographic limitations in LASV surveillance, particularly, in non-human hosts. Further \n47 investigation of the non-human reservoir of LASV, alongside expanded surveillance, are \n48 required for precise characterisation of the emergence and dispersal of LASV. Accurate \n49 surveillance of LASV circulation in non-human hosts is vital to guide early detection and \n50 initiation of public health interventions for future Lassa fever outbreaks. \n51\n52 Key-words\n53 Lassa mammarenavirus; Lassa Fever; Phylogeography; Metadata; Zoonoses; Surveillance\n54\n55\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 3 of 18\n56 1 Introduction\n57\n58 Lassa fever (LF) is a lethal zoonotic viral haemorrhagic disease of humans, caused by Lassa \n59 mammarenavirus (LASV). It causes an estimated 900,000 annual human infections and \n60 several thousand deaths in West Africa annually (1,2). The WHO assigns LASV endemicity to \n61 eight West African countries: Benin, Ghana, Guinea, Liberia, Mali, Sierra Leone, Togo and \n62 Nigeria (S1 Fig) (3). LASV is a bisegmented ssRNA- virus of the family Arenaviridae (4,5). \n63 Based on the genomic analysis of the large (L) and small segments (S) LASV has been \n64 classified into seven lineages which demonstrate spatial segregation across the endemic \n65 range (6). The high nucleotide variability (25-32%) of these lineages introduces complexity \n66 into assays to detect LASV infection.\n67\n68 Epidemiological data on LF is limited and constrained by current testing and reporting in the \n69 endemic region, making accurate estimates of its true burden challenging (7). Many individuals \n70 infected with LASV do not seek healthcare with up to 80% of infections assumed \n71 asymptomatic or presenting as mild illness (8). Estimates based on longitudinal serological \n72 surveys in Sierra Leone in the early 1980’s indicated that 100,000 to 300,000 infections of LF \n73 occurred annually in West Africa, with more recent estimates being up to 900,000 infections \n74 (2,8). Identification of symptomatic cases is further confounded by overlapping symptoms with \n75 other diseases (e.g., malaria) and lack of available diagnostic methods (1,9–11). Access to \n76 diagnostic tests varies spatially, increased availability at centers of excellence in LF treatment \n77 and research such as the Irrua Specialist Teaching Hospital, Nigeria and Kenema General \n78 Hospital, Sierra Leone results in a spatial bias of reported cases from these locations. \n79 Phylogenetic analysis and molecular dating of sequence clinical and research samples \n80 suggest a westward route of dispersal of LASV lineages, from the most recent common \n81 ancestor in Nigeria. (12–18). These estimates have been used to project the potential for \n82 Lassa Fever to extend beyond the current endemic zone (19).\n83\n84 The Natal multimammate mouse (Mastomys natalensis) is the primary reservoir of LASV, \n85 however, 11 other rodent species have been found to be acutely infected or have seropositivity \n86 to LASV including; Mastomys erythroleucus, Hylomyscus pamfi, Mus baoulei and Rattus \n87 rattus (15,20–24). Humans become infected with LASV upon contact with or inhalation of \n88 excretions from the rodent species (12,25). Although human-to-human transmission has been \n89 reported – typically associated with nosocomial outbreaks – these are rare events when \n90 compared with spillover from rodent hosts  (26). We performed a study of LASV nucleotide \n91 sequences available from the National Centre for Biotechnology Information (NCBI) GenBank, \n92 using associated metadata to spatially model sequencing effort, adjusted for the number of \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 4 of 18\n93 suspected and confirmed human LF cases to determine potential biases in locations of \n94 available sequences or significant geographic limitations in LASV surveillance, particularly, in \n95 non-human hosts. \n96   \n97\n98 2 Methods\n99\n100 2.1 Data Collection and Processing\n101\n102 LASV nucleotide and protein sequences were obtained from the NCBI GenBank (27). The \n103 search query run on 24 Sep 2021 was for “Lassa mammarenavirus” in the organism field of \n104 the NCBI nucleotide dataset. Data were obtained using the NCBI Entrez API with analysis \n105 conducted using the “genbankr” package within the R statistical programming language (27–\n106 29). Associated citations were manually retrieved to identify missing metadata for sequences \n107 including hosts and geographic location of samples. Sequences with large portions (10% \n108 missing compared to reference sequences, NC_004296.1 and NC_004297.1 for S and L \n109 segments respectively) of missing nucleotide data on the L- or S-segment or lacking \n110 associated metadata (collection year, host species, country, and geographical region of \n111 sampling) were excluded from phylogenetic analysis. Nucleotide sequences were aligned \n112 using the ‘map to reference’ tool on Geneious Prime 20201.2. Alignment, visual inspection \n113 and manual editing were performed, and entries that contained >100 continuous ambiguous \n114 nucleotide calls were excluded (S1 Data). \n115\n116 2.2 Sequencing Bias\n117\n118 First, we compared the number of cases reported from countries between 2008-2023 with the \n119 number of samples contained in GenBank to summarise the correlation between reported \n120 human cases and availability of sequences. We then compared the proportion of human to \n121 non-human derived sequences within countries. \n122\n123 To understand the bias of sequenced samples at a sub-national level the origin of a sequenced \n124 sample was geocoded using the Google Geocoding API using the “ggmap” package (30). \n125 Sequence locations were associated with level-1 administrative regions and data were \n126 separated into human and rodent sources of samples to visualise the spatial heterogeneity of \n127 sampling. To measure sampling effort bias, the number of samples obtained within a level-1 \n128 administrative region was associated with the centroid of the region. The number of confirmed \n129 LF clinical cases reported from these regions in the previous 15 years was obtained (S2 Data). \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 5 of 18\n130 The number of cases within a region was divided by the human population count to produce \n131 the number of confirmed cases per 100,000 individuals. The number of sequences was used \n132 as the response variable in a spatial Generalised Additive Model, with geographic coordinates \n133 and cases per 100,000 individuals used as covariates. This model was constructed using the \n134 “mgcv” package (31).\n135\n136 2.3 Phylogenetic Analysis\n137\n138 Phylogenetic analysis was undertaken through Bayesian Markov Chain Monte Carlo (MCMC) \n139 method using BEAST.v1.10.4 (32). In BEAUTi, the parameters were a substitution model as \n140 a generalised time reversible plus gamma site heterogeneity, with codon partition positions 1, \n141 2, 3. A strict clock and a coalescent tree prior with a constant size population was used. Each \n142 analysis consisted of 20 million MCMC steps and trees were sampled every 20,000 \n143 generations. Sample collection dates from the metadata were used as tip dates to fit to a \n144 molecular clock, and country of sample collection was incorporated as a discrete state (16,33). \n145 To assess the log files of the output TRACER.v.1.7.1 was used. Maximum-clade credibility \n146 trees were generated through TreeAnnotator v1.8.4 and visualised in FigTree.v1.4.4 (34). \n147\n148 3 Results\n149\n150 3.1 Compiled Dataset\n151\n152 The initial dataset comprised 2,298 records (from samples obtained 1969-2019), including \n153 nucleotide sequences and associated metadata. Incomplete gene sequences and sequences \n154 lacking metadata information (n = 1,045) were removed from phylogenetic analyses. \n155 Therefore, 680 sequences of complete S segment and 573 sequences of partial L segment (L \n156 protein only) were used. Accession numbers of included and excluded sequences are \n157 available in S1 Data.\n158\n159 3.2 Descriptive Analysis\n160\n161 Year of collection was available for 2,108 records, with the oldest sequence dating from 1969 \n162 and latest from 2019. Among these records, most sequences (n = 1,936, 92%) have been \n163 obtained since 2008. Human-derived LASV sequences comprised most of the available \n164 records (67%), other host species include Mastomys natalensis (29%) and Mastomys spp. \n165 (3%), while Mastomys erythroleucus (n = 18), Mus baoulei (n = 9) and Hylomyscus pamfi (n = \n166 10) represent < 1% each. The species sampled was not documented in 107 records. Country \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 6 of 18\n167 of collection was available for 2,238 records. Most sequences were produced from samples \n168 collected in Nigeria (54%), followed by Guinea (20%), Sierra Leone (14%), Liberia (4%) and \n169 Cote d’Ivoire (3%) with the remainder obtained from, Benin, Ghana, Mali and Togo (Fig 1).\n170\n171 Sequences for human derived samples with regional location data (n = 1328, 63%) were \n172 clustered in Edo State, Nigeria (n = 519, 39%), Ondo State, Nigeria (n = 220, 17%) and \n173 Eastern Province, Sierra Leone (n = 159, 12%) with 430 samples from the remaining endemic \n174 regions. Sequences from rodent samples with regional location data (n = 527, 25%) were most \n175 commonly obtained from Faranah, Guinea (n = 210, 39%) and Eastern Province, Sierra Leone \n176 (n = 107, 20%) with 210 samples from the regions. \n177\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 7 of 18\n178\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 8 of 18\n179 Figure 1 – The number of sequences, shown on a log 10 scale, retrieved from NCBI GenBank \n180 with associated regional sampling location and host for human samples (top, n = 1,328) and \n181 rodent samples (bottom, n = 527). Grey regions represent level-1 administrative areas with no \n182 sequences within countries that have at least one available sequence. White countries are \n183 West African countries with no available LASV sequences. See S1 Fig for country names. \n184 Shapefiles for basemap layer obtained from GADM 4.0.2 (35)\n185\n186\n187 3.3 Sequencing bias\n188\n189 We observed a strong positive correlation between the number of confirmed human cases \n190 between 2008-2023 and the number of GenBank deposited sequences at country level \n191 (r(degrees of freedom = 7) = 0.93 (95% Confidence Interval = 0.71-0.98), p < 0.001). When \n192 analysed by species source no correlation was observed with the number of confirmed cases \n193 and the number of available rodent sequences was observed (r(6) = -0.019 (95% C.I. -0.71-\n194 0.69), p = 0.96). \n195\n196 When combining both human and rodent-derived samples at the regional level to explore \n197 spatial sampling biases, we found that sequencing effort is greatest in Southwest Nigeria, \n198 centred over Edo State and the Faranah and Nzérékoré regions of Guinea, Eastern Province \n199 of Sierra Leone and Nimba district of Liberia (Fig 2). There was a positive, non-linear \n200 association between the rate of confirmed human cases with the number of available rodent \n201 and human derived LASV sequences at regional level (deviance explained = 14%, estimated \n202 degrees of freedom = 2.29, p < 0.001). \n203\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 9 of 18\n204\n205 Figure 2 – Modelled relative sequencing effort derived from both human and rodent samples. \n206 Greatest sequencing effort coincides with areas where sampling in humans (Edo, Nigeria and \n207 Kenema, Sierra Leone) and rodents (Faranah, Guinea) have historically been focussed.  \n208 Shapefiles for basemap layer obtained from GADM 4.0.2 (35)\n209\n210 3.4 Phylogenetic Analysis\n211\n212 Sequences for each segment of LASV showed clustering according to previously documented \n213 lineages I-VII alongside geographical clustering with lineages I-III and VI present in Nigeria, \n214 lV in Liberia, Guinea and Sierra Leone, V in Mali and VII in Togo (S2 Fig). In this analysis only \n215 L segment sequences of lineage V from Cote d’Ivoire were included due to quality control \n216 exclusion criteria. The phylogeny of the L segment indicates an older emergence of LASV in \n217 the human population, with the most recent common ancestor (MRCA) predicted in the year \n218 828 in Nigeria, inference based on the S segment indicates the emergence in the year 1350 \n219 (Table 1). \n220\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 10 of 18\n221 Table 1 - The most recent common ancestor (MRCA) stratified by host and country of \n222 collection of Lassa mammarenavirus (LASV) S and L segments. Samples were collected \n223 between 1969-2018.\nHost species Country S segment MRCA L segment MRCA\nBenin 1995 1989\nGuinea 1895 1871\nLiberia 1895 1627\nNigeria 1681 1498\nSierra Leone 1901 1874\nHomo sapiens (n=1181)\nTogo 2016 2014\nHylomyscus pamfi (n=2) Nigeria 1681 1498\nGuinea 1975 2010Mastomys erythroleucus \n(n=18) Nigeria 2008 2006\nGuinea 1938 1997\nMali 1951 2007\nMastomys natalensis \n(n=36)\nSierra Leone 1909 1979\n224\n225 There was a lack of sequence information from lineage I and VI, however, phylogeny suggests \n226 these lineages are basal to others in Nigeria (S2 Fig). Lineage VII in Togo is most closely \n227 related to Nigerian isolates and potentially diverged between 500-900 years ago. The \n228 divergence of lineage III and IV is predicted to have occurred between the years 1332-1551. \n229 Introduction to countries west of Nigeria appears to be by dispersal initially to Liberia, followed \n230 by Guinea in the 1700s, followed by Sierra Leone and Mali approximately 100 years later. A \n231 lack of full segment sequences from lineage V limits calculation of divergence from the most \n232 recent common ancestor from lineage IV (approximately 200 years). \n233\n234\n235 4 Discussion\n236\n237 There are several important aspects of our study and findings. First, we studied a \n238 comprehensive dataset of publicly available full-segment LASV sequences, spanning West \n239 Africa and host species, to inform our understanding of the phylogeny of LASV dispersal. \n240 Second, we identified substantial variability in the origin of available sequences and \n241 completeness of records. Third, we showed strong geographic clustering among lineages \n242 supporting prior hypotheses of radiation from both Nigeria and a subsequent introduction into \n243 Liberia (19). Fourth, the synthesis of available metadata highlights important gaps in currently \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 11 of 18\n244 available data, including spatial bias in the sequencing of samples and suggests this should \n245 be used to inform the design of epidemiological programmes going forward. \n246\n247 Our analyses of 2,298 LASV sequences obtained from GenBank highlights the spatial biases \n248 in the availability of sequence data that may limit our understanding of the current and historic \n249 dispersal of LASV lineages in West Africa.  First, sequence data was typically obtained from \n250 three of the eight endemic countries: Nigeria, Guinea and Sierra Leone. We found a strong \n251 associated between the number of reported human cases and number of available sequences. \n252 When stratifying by host species this trend did not remain with rodent derived samples \n253 showing no association with the number of human cases indicating important under-sampling \n254 in high human cases regions and relatively high sampling in locations with low numbers of \n255 human cases. This is potentially an important source of bias when attempting to infer \n256 phylogeography within the reservoir host of this zoonotic pathogen. Sequence data from other \n257 countries, and more regions within them, across West Africa are required to increase \n258 confidence in the timelines of the currently inferred westward expansion. Greater focus needs \n259 to be placed on acquiring sequences from the rodent host to understand viral genetic diversity \n260 within the primary reservoir species. Comparing rodent derived sequences with those \n261 acquired from spillover into human populations may also allow identification of genetic drivers \n262 of transmission (36).\n263\n264 The overrepresentation of data from these three countries has been mapped as relative \n265 sequencing effort to identify regions where increased LASV sequencing are required to \n266 counteract current sequencing biases. Second, geographic clustering of LASV lineages, \n267 suggest isolated events of human-to-rodent transmission and the emergence of LASV dating \n268 from 1498 in Nigeria. Similarly, Olayemi et al. report evidence of earlier emergence of the virus \n269 in humans than in rodents in Nigeria (16). Comparatively limited data from non-human hosts \n270 with limited genome coverage, (69/703 sequences encompassed complete genes) produce \n271 important uncertainty around the observation of human-to-rodent transmission. Taken \n272 together, this data highlight limited surveillance among animal species, necessitating further \n273 investments in data acquisition and sharing to accurately define the spatiotemporal expansion \n274 of LASV in West Africa. \n275\n276 The phylogenetic analysis of LASV stratified by host species supports spatial evolution, in \n277 addition to intra-host viral evolution (S2 Fig). For instance, LASV sequences from M. \n278 erytholeucus sampled in Nigeria and Guinea clustered within lineages III and IV, respectively. \n279 Interestingly, these isolates appear to occur after the emergence of the most recent common \n280 ancestor virus circulating among humans and M. natalensis in these countries (Table 1), \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 12 of 18\n281 suggesting introduction of LASV into M. erythroleucus populations was a consequence of \n282 pathogen circulation in human and M. natalensis populations. Sequences from M. natalensis \n283 in Sierra Leone exhibit minimal clustering, and were interspersed with sequences from \n284 humans, potentially representing isolated events of pathogen introduction into human \n285 populations with spillback into commensal rodent populations (i.e., reverse zoonosis). The \n286 most recent common ancestor of LASV sequences from M. natalensis in Sierra Leone suggest \n287 a later emergence of the virus in this country. Our findings corroborate those of Olayemi et al., \n288 that within Sierra Leone LASV appears to have emerged in human hosts before rodents (16). \n289 However, this data must be caveated by the limited information from rodent species in these \n290 locations.\n291\n292 There is a lower coverage of rodent-derived LASV sequences, with those from the primary \n293 reservoir M. natalensis forming fewer than one-third of all sequences (n = 642, 28%), with \n294 substantially lower sampling of other possible rodent hosts, including other Mastomys species. \n295 Rodent sampling has not increased at the same rate as human samples despite increased \n296 sampling effort since 2008 (15,22,37). There is substantial heterogeneity in the locations in \n297 which rodent and human samples are available. For example, a relatively high number of \n298 rodent samples (n = 429) have been obtained from Guinea while few human sequences (n = \n299 20) are available from these locations. The inverse is true of Nigeria where most human \n300 derived sequences are obtained (n = 1,147) but only 85 rodent sequences are available, and \n301 all of these from a single state (Edo, Nigeria). The number of suspected and reported cases \n302 was found to be positively but non-linearly associated with the number of available sequences. \n303 This is suggestive of a consolidation of research and focus of sampling in areas historically \n304 with high numbers of human cases but has led to a paucity of sequences from elsewhere in \n305 the endemic region. The limited number of full segment sequences from rodents, from few \n306 geographic locations, limits our understanding of viral radiation in rodent hosts, particularly \n307 from species which are not considered the primary reservoir, e.g., H. pamfi. The most recent \n308 common ancestor for the viral sequence obtained from H. pamfi is estimated to be in the late \n309 1600s, it is therefore possible lineage VI and/or H. pamfi as a reservoir of LASV has gone \n310 undetected due to lack of sufficient sampling (15). \n311\n312 Interpreting available LASV sequences is challenging for several reasons. A large proportion \n313 of available sequences (70%) have been obtained within Lassa fever research programs, \n314 representing spatial ascertainment bias (38–40). In addition to these spatial biases’ temporal \n315 biases are apparent. Since 2016 there has been a substantial increase in the number of LASV \n316 sequences available in NCBI GenBank, reflecting increasing research effort, availability of \n317 sequencing platforms and increased data collection during Lassa fever epidemics, such as in \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 13 of 18\n318 the 2018 Nigeria Lassa fever outbreak (41–43). There are notably fewer recorded sequences \n319 of LASV from Benin, Togo, and Ghana, suggesting a potential a gap in surveillance and \n320 research capacity in these locations or a lack of circulating LASV, despite several reported \n321 outbreaks (44–46). Phylogenetic analysis on 60% of our initial dataset, following removal of \n322 sequences due to incompleteness or missing geographic and year of collection information (n \n323 = 1,045) demonstrated geographic clustering of LASV lineages, supporting prior analyses \n324 (14–16,33,44,47–49). Increased data availability from Nigeria following increased LASV \n325 surveillance allowed regional analysis of phylogeny for lineages II and III supporting previous \n326 findings of expansion of these lineages from North-East Nigeria to the South-West of the \n327 country (13,50,51). \n328\n329 A substantial number (n = 869) of the sequences retrieved corresponded to short fragments \n330 (< 1 Kb) probably derived from PCR products used for diagnostic purposes rather than for \n331 viral genomic surveillance. LASV is a segmented virus, and it was not possible to identify \n332 complete genome sequences since both S and L segments are reported separately on the \n333 sequence’s repository. The molecular clock analyses from L protein indicated an earlier \n334 emergence of LASV when compared to S segment analysis (828 and 1350 respectively), \n335 potentially because the viral RNA polymerase (L protein) is less affected by selective \n336 pressure than the S segment (12,47,52).\n337\n338 Despite these challenges, this study has synthesised currently available data on LASV \n339 sequences to investigate the location and period of sampling to reconstruct the dispersal of \n340 viral lineages across the endemic region. Despite the regionalisation of LF being driven by \n341 rodent-to-human transmission, there remains scarce LASV genomic data from non-human \n342 hosts. We have mapped the locations of relative under sampling to guide targeted efforts to \n343 counteract biases in currently available data for both rodent and human derived sequences. \n344 Expanded sampling of LASV from animal species within the endemic region will improve our \n345 current understanding of LASV evolution and ecology and improve confidence in current \n346 estimates of westward expansion of Lassa fever in humans. Further understanding of the \n347 viral evolution dynamics of LASV and spatial expansion of current lineages will be vital to \n348 ensure adequate diagnostic tools are available to respond to the expected sporadic \n349 outbreaks of Lassa Fever across the region.\n350\n351 Supplementary material\n352\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 14 of 18\n353 S1 Data. GenBank accession number of analysed sequences. This dataset includes \n354 available data about host, country, region, year, sequence length, genome segment (L or S) \n355 and predicted MRCA.\n356\n357 S2 Data. Dataset on confirmed Lassa fever cases. This presents the number of confirmed \n358 cases of Lassa fever reported from countries between 2008 and 2023 at a subnational level \n359 that were used to calculate the number of cases per 100,000 people. References for the \n360 reports used to produce this dataset are included.\n361\n362 S1 Figure. Map of West Africa.  displays a map of West Africa with country names for \n363 reference with Fig 1 and Fig 2. Shapefiles for mapping obtained from GADM 4.0.2 (35)\n364\n365 S2 Figure. Time-calibrated phylogeny for both the small segment (S) and large segment \n366 (L) from included LASV sequences.\n367\n368 Author contributions\n369 Conceptualisation: DS and LBA; Methodology: HF, DS, DA and LBA; Formal Analyses: HF, \n370 DS and LBA; Investigation: HF, DS and LBA; Supervision: LBA; Data Curation: HF and DS; \n371 Writing – original draft preparation: HF, DS, LBA; Writing – Review and Editing: IH, LE, NH, \n372 RA, RK, FN, DA, AZ and TMcH; Funding acquisition: AZ and FN.\n373\n374 Data availability and reproducibility\n375 All data used in these analyses are publicly available from GenBank. The accession numbers \n376 of records used are available as supplementary material. Code to reproduce the metadata \n377 analyses are available as an archived Git release on Zenodo \n378 (https://doi.org/10.5281/zenodo.6340162)\n379\n380 Conflict of interests\n381 The authors declare no conflict of interests\n382\n383 Acknowledgements:\n384 Linzy Elton, Timothy D McHugh, Francine Ntoumi, and Alimuddin Zumla acknowledge support \n385 from EDCTP-Central Africa and East African Clinical Research Networks (CANTAM-3, \n386 EACCR-3).  Sir Zumla is an NIHR Senior Investigator, a Mahathir Science Award, Sir Patrick \n387 Manson Medal and EU-EDCTP Pascoal Mocumbi Prize laureate. Liã Bárbara Arruda, David \n388 Simons, Rashid Ansumana, Linzy Elton, Najmul Haider, Isobella Honeyborne, Danny Asogun, \n389 Timothy D McHugh, Francine Ntoumi, Alimuddin Zumla and Richard Kock acknowledge \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 15 of 18\n390 support from the Pan-African Network for Rapid Research, Response and Preparedness for \n391 Infectious Diseases Epidemics – PANDORA-ID-NET, funded through the European and \n392 Developing Countries Clinical Trials Partnership (EDCTP) (grant number RIA2016E-1609). \n393 David Simons is supported by a PhD studentship from the UK Biotechnology and Biological \n394 Sciences Research Council (BB/M009513/1).\n395\n396 References\n397 1. Asogun DA, Gunther S, Akpede GO, Ihekweazu C, Zumla A. Lassa Fever: Epidemiology, Clinical \n398 Features, Diagnosis, Management and Prevention. [Review]. Infectious Disease Clinics of North \n399 America. 2019;33(4):933–51. \n400 2. Basinski AJ, Fichet-Calvet E, Sjodin AR, Varrelman TJ, Remien CH, Layman NC, et al. Bridging the \n401 gap: Using reservoir ecology and human serosurveys to estimate Lassa virus spillover in West \n402 Africa. Wesolowski A, editor. PLoS Comput Biol. 2021 Mar 3;17(3):e1008811. \n403 3. World Health Organisation. Lassa fever [Internet]. 2022 [cited 2022 Feb 22]. Available from: \n404 https://www.who.int/health-topics/lassa-fever#tab=tab_1\n405 4. Gunther S, Lenz O. Lassa virus. [Review] [323 refs]. Critical Reviews in Clinical Laboratory \n406 Sciences. 2004;41(4):339–90. \n407 5. Hallam SJ, Koma T, Maruyama J, Paessler S. Review of Mammarenavirus Biology and Replication. \n408 Front Microbiol [Internet]. 2018 [cited 2020 Oct 21];9. Available from: \n409 https://www.frontiersin.org/articles/10.3389/fmicb.2018.01751/full\n410 6. Welch SR, Scholte FEM, Albariño CG, Kainulainen MH, Coleman-McCray JD, Guerrero LW, et al. \n411 The S Genome Segment Is Sufficient to Maintain Pathogenicity in Intra-Clade Lassa Virus \n412 Reassortants in a Guinea Pig Model. Frontiers in Cellular and Infection Microbiology [Internet]. \n413 2018 [cited 2022 Feb 3];8. Available from: \n414 https://www.frontiersin.org/article/10.3389/fcimb.2018.00240\n415 7. Simons D. Lassa fever cases suffer from severe underreporting based on reported fatalities. \n416 International Health. 2022; \n417 8. McCormick JB, Webb PA, Krebs JW, Johnson KM, Smith ES. A prospective study of the \n418 epidemiology and ecology of Lassa fever. J Infect Dis. 1987;155(3):437–44. \n419 9. Takah NF, Brangel P, Shrestha P, Peeling R. Sensitivity and specificity of diagnostic tests for Lassa \n420 fever: a systematic review. BMC Infectious Diseases. 2019 Jul 19;19(1):647. \n421 10. Nnaji ND, Onyeaka H, Reuben RC, Uwishema O, Olovo CV, Anyogu A. The deuce-ace of Lassa \n422 Fever, Ebola virus disease and COVID-19 simultaneous infections and epidemics in West Africa: \n423 clinical and public health implications. Tropical Medicine and Health. 2021 Dec 30;49(1):102. \n424 11. Ashcroft JW, Olayinka A, Ndodo N, Lewandowski K, Curran MD, Nwafor CD, et al. Pathogens that \n425 Cause Illness Clinically Indistinguishable from Lassa Fever, Nigeria, 2018. Emerging Infectious \n426 Diseases. 2022;28(5):994–7. \n427 12. Andersen KG, Shapiro BJ, Matranga CB, Sealfon R, Lin AE, Moses LM, et al. Clinical Sequencing \n428 Uncovers Origins and Evolution of Lassa Virus. Cell. 2015; \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 16 of 18\n429 13. Bowen MD, Rollin PE, Ksiazek TG, Hustad HL, Bausch DG, Demby AH, et al. Genetic Diversity \n430 among Lassa Virus Strains. Journal of Virology. 2000;74(15):6992–7004. \n431 14. Manning JT, Forrester N, Paessler S. Lassa virus isolates from Mali and the Ivory Coast represent \n432 an emerging fifth lineage. Frontiers in Microbiology. 2015; \n433 15. Olayemi A, Cadar D, Magassouba N, Obadare A, Kourouma F, Oyeyiola A, et al. New Hosts of The \n434 Lassa Virus. Scientific Reports. 2016; \n435 16. Olayemi A, Adesina AS, Strecker T, Magassouba N, Fichet-Calvet E. Determining ancestry \n436 between rodent-and human-derived virus sequences in endemic foci: Towards a more integral \n437 molecular epidemiology of lassa fever within West Africa. Biology. 2020; \n438 17. Whitmer SLM, Strecker T, Cadar D, Dienes HP, Faber K, Patel K, et al. New lineage of lassa virus, \n439 Togo, 2016. Emerging Infectious Diseases. 2018;24(3):599–602. \n440 18. Okoro OA, Bamgboye E, Dan-Nwafor C, Umeokonkwo C, Ilori E, Yashe R, et al. Descriptive \n441 epidemiology of Lassa fever in Nigeria, 2012-2017. Pan Afr Med J. 2020 Sep 3;37:15. \n442 19. Klitting R, Kafetzopoulou LE, Thiery W, Dudas G, Gryseels S, Kotamarthi A, et al. Predicting the \n443 evolution of Lassa Virus endemic area and population at risk over the next decades [Internet]. \n444 Microbiology; 2021 Sep [cited 2022 Feb 3]. Available from: \n445 http://biorxiv.org/lookup/doi/10.1101/2021.09.22.461380\n446 20. Bangura U, Buanie J, Lamin J, Davis C, Bongo GN, Dawson M, et al. Lassa Virus Circulation in \n447 Small Mammal Populations in Bo District, Sierra Leone [Internet]. Vol. 10, BIOLOGY-BASEL. ST \n448 ALBAN-ANLAGE 66, CH-4052 BASEL, SWITZERLAND: MDPI; 2021. Available from: \n449 https://doi.org/10.3390/biology10010028\n450 21. Forni D, Sironi M. Population Structure of Lassa Mammarenavirus in West Africa. Viruses. \n451 2020;12(4):437. \n452 22. Lecompte E, Fichet-Calvet E, Daffis S, Koulémou K, Sylla O, Kourouma F, et al. Mastomys \n453 natalensis and Lassa fever, West Africa. Emerging Infectious Diseases. 2006; \n454 23. Wulff H, Fabiyi A, Monath TP. Recent isolations of Lassa virus from Nigerian rodents. Bull World \n455 Health Organ. 1975;52(4–6):609–13. \n456 24. Yadouleton A, Agolinou A, Kourouma F, Saizonou R, Pahlmann M, Bedié SK, et al. Lassa virus in \n457 pygmy mice, Benin, 2016-2017. Emerging Infectious Diseases. 2019; \n458 25. Oti VB. A Reemerging Lassa Virus: Aspects of Its Structure, Replication, Pathogenicity and \n459 Diagnosis. In: Alfonso J. Rodriguez-Morales, editor. Current Topics in Tropical Emerging Diseases \n460 and Travel Medicine. BoD – Books on Demand; 2018. \n461 26. Lo Iacono G, Cunningham AA, Fichet-Calvet E, Garry RF, Grant DS, Khan SH, et al. Using \n462 Modelling to Disentangle the Relative Contributions of Zoonotic and Anthroponotic \n463 Transmission: The Case of Lassa Fever. PLoS Neglected Tropical Diseases. 2015; \n464 27. National Center for Biotechnology Information. National Center for Biotechnology Information \n465 [Internet]. 2022 [cited 2022 Feb 3]. Available from: https://www.ncbi.nlm.nih.gov/\n466 28. Becker G, Lawrence M. genbankr: Parsing GenBank files into semantically useful objects. 2021. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 17 of 18\n467 29. R Core Team. R: A Language and Environment for Statistical Computing [Internet]. Vienna, \n468 Austria: R Foundation for Statistical Computing; 2021. Available from: https://www.R-\n469 project.org/\n470 30. Kahle D, Wickham H. ggmap: Spatial Visualization with ggplot2. The R Journal. 2013;5(1):144–61. \n471 31. Wood SN. Generalized Additive Models: An Introduction with R. 2nd ed. Chapman and Hall/CRC; \n472 2017. \n473 32. Suchard MA, Lemey P, Baele G, Ayres DL, Drummond AJ, Rambaut A. Bayesian phylogenetic and \n474 phylodynamic data integration using BEAST 1.10. Virus Evolution. 2018; \n475 33. Olayemi A, Fichet-Calvet E. Systematics, ecology, and host switching: Attributes affecting \n476 emergence of the Lassa virus in rodents across western Africa. Viruses. 2020. \n477 34. Rambaut A, Drummond AJ, Xie D, Baele G, Suchard MA. Posterior summarization in Bayesian \n478 phylogenetics using Tracer 1.7. Systematic Biology. 2018; \n479 35. Database of Global Administrative Areas. GADM [Internet]. 2022 [cited 2021 Apr 25]. Available \n480 from: https://gadm.org/index.html\n481 36. Whitlock AOB, Bird BH, Ghersi B, Davison AJ, Hughes J, Nichols J, et al. Identifying the genetic \n482 basis of viral spillover using Lassa virus as a test case. R Soc Open Sci. 2023 Mar 22;10(3):221503. \n483 37. Lecompte E, Brouat C, Duplantier JM, Galan M, Granjon L, Loiseau A, et al. Molecular \n484 identification of four cryptic species of Mastomys (Rodentia, Murinae). Biochemical Systematics \n485 and Ecology. 2005; \n486 38. Townsend Peterson A, Moses LM, Bausch DG. Mapping transmission risk of lassa fever in West \n487 Africa: The importance of quality control, sampling bias, and error weighting. PLoS ONE. 2014; \n488 39. Ehichioya DU, Hass M, Ölschläger S, Becker-Ziaja B, Onyebuchi Chukwu CO, Coker J, et al. Lassa \n489 fever, Nigeria, 2005-2008. Emerging Infectious Diseases. 2010. \n490 40. Khan SH, Goba A, Chu M, Roth C, Healing T, Marx A, et al. New opportunities for field research \n491 on the pathogenesis and treatment of Lassa fever. Antiviral Research. 2008; \n492 41. Maxmen A. Deadly Lassa-fever outbreak tests Nigeria’s revamped health agency. Nature. \n493 2018;555(7697):421–2. \n494 42. Siddle KJ, Eromon P, Barnes KG, Mehta S, Oguzie JU, Odia I, et al. Genomic Analysis of Lassa \n495 Virus during an Increase in Cases in Nigeria in 2018. New England Journal of Medicine. 2018 Nov \n496 1;379(18):1745–53. \n497 43. Ilori EA, Frank C, Dan-Nwafor CC, Ipadeola O, Krings A, Ukponu W, et al. Increase in Lassa Fever \n498 Cases in Nigeria, January–March 2018. Emerging Infectious Diseases [Internet]. 2019 May [cited \n499 2020 Oct 21];25(5). Available from: 10.3201/eid2505.181247\n500 44. Yadouleton A, Picard C, Rieger T, Loko F, Cadar D, Kouthon EC, et al. Lassa fever in Benin: \n501 description of the 2014 and 2016 epidemics and genetic characterization of a new Lassa virus. \n502 Emerging Microbes & Infections. 2020;1–23. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\nPage 18 of 18\n503 45. World Health Organisation. Lassa Fever – Togo [Internet]. 2022 [cited 2022 Nov 24]. Available \n504 from: https://www.who.int/emergencies/disease-outbreak-news/item/2022-DON362\n505 46. Ghana Health Services. Lassa Fever Press Release Ghana 2023. 2023 Apr 19 [cited 2023 Apr 19]; \n506 Available from: https://osf.io/ft2gy/\n507 47. Ibukun FI. Inter-lineage variation of lassa virus glycoprotein epitopes: A challenge to lassa virus \n508 vaccine development. Viruses. 2020. \n509 48. Lalis A, Leblois R, Lecompte E, Denys C, ter Meulen J, Wirth T. The Impact of Human Conflict on \n510 the Genetics of Mastomys natalensis and Lassa Virus in West Africa. PLoS ONE. 2013;7(5). \n511 49. Wiley MR, Fakoli L, Letizia AG, Welch SR, Ladner JT, Prieto K, et al. Lassa virus circulating in \n512 Liberia: a retrospective genomic characterisation. The Lancet Infectious Diseases. 2019; \n513 50. Ehichioya DU, Hass M, Becker-Ziaja B, Ehimuan J, Asogun DA, Fichet-Calvet E, et al. Current \n514 molecular epidemiology of Lassa virus in Nigeria. Journal of Clinical Microbiology. 2011; \n515 51. Naidoo D, Ihekweazu C. Nigeria’s efforts to strengthen laboratory diagnostics – Why access to \n516 reliable and affordable diagnostics is key to building resilient laboratory systems. African Journal \n517 of Laboratory Medicine [Internet]. 2020 Aug 26 [cited 2020 Oct 21];9(2). Available from: \n518 10.4102/ajlm.v9i2.1019\n519 52. Hastie KM, Saphire EO. Lassa virus glycoprotein: stopping a moving target. Current Opinion in \n520 Virology. 2018. \n521\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint \n\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprintthis version posted June 22, 2023. ; https://doi.org/10.1101/2023.06.20.23291686doi: medRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}