2D.BENVENUTO ET AL nlm.nih.gov/genbank/and GISAID (https://w v.gisaid substantial temporalsignaland the possibility tocalibrate ase a reliabl despite the limit progr )Th was perfo ar clock, dal den anhic model with an uncorrelated relaxed cock TREE 1.6.8 software (http://www.igtree.org)[8] as the one that best fit the data.Molecular clock calibra A maximum (ML)phylogeny was estimated the evolu ary rate of the 2019-nCov unde the HK e geno s the MCC tree (https//qithub.com/ddarriba/imodeltest2)as the best fit- ting model [9] able origin of 2019-nCoV In order stigate the tempor signal,we state posteri regresse f the 15.1 (http://101.The ML phvlo cA of the 95%6HPD was used as a starting tree for Bayesian time-scaled while the MRCA of Bat SARS-like Coronavirus and related phylogeneti c analysis using BEAST 1.10.4(http: Deas 2019-nCoV lineages dates back to February 22.201 er20,2 ugus 14)(Fig and stepping stone (SS)pr dures to estimate the most appropriate molecular clock model for the The first evidence of 2019-nCoV dissemination app ars to Bayesian phylogenetic analysis [12].We tested a)the be,acc ding to our phylogeographic reconstruction strict molecula ock mo which assur a single rom Wuhar hina,to N nthamburi,Tha vith a spp.or u.s by the ith fur th with a logpormal rate distribution (Uc N)31 Both 55 the second one following a more complex patter:from and PS estimators indicated the uncorrelated relaxed Nonthamburi to Zhejiang,Huangzhou (spp =0.47).as r clock(Bayes Factor 4.3)as the be 0=0.62 analysis. Zh The stitution model and the B red to be linke model of population size and growth [14].We com to Guandona Zhuhai isolates in agreement with report puted MCMC (Markov chair Monte Carlo)duplicate of patients trav veling back from that region of China mpling s the most pro of the f the 2019-nCoV (spp 200 for eac ated nar Discussion was obtaine the distr is know about 2019-nCoV virus,indluding (http:/ 09 fter 10%6 burn-in ce 05% the past few months of the ongoing epidemic outbreak is not neces Results arily surprising for an RNA virus that has been shown to Despite sinc the beginning of the ep e ably evol ng populat ve ort tim s unders th urgen in15%of the sites11%of which were pa ony infor ment of appr indepth monitoring sste mative,thus indicating the presence of sufficient phylo capable of investigating viral mutation and transmission genetic sign al for further analysis,in agree with the as 2019-nCov unt ey keeps spreading level by and al le the full dataset show d hiah correlation hetv ence of novel transmission pling time and genetic distance to the root of the ML tree routes and/or patterns should be considered a significant of the available sequences (R-squared 0.85),indicating priority.nlm.nih.gov/genbank/) and GISAID (https://www.gisaid. org/) databases. Alignment was performed using MAFFT online program [7]. The complete dataset was assessed for presence of phylogenetic signal by applying the likelihood mapping analysis implemented in the IQ￾TREE 1.6.8 software (http://www.iqtree.org) [8]. A maximum likelihood (ML) phylogeny was recon￾structed using IQ-TREE 1.6.8 software under the HKY nucleotide substitution model with four gamma cate￾gories (HKY+G4), which was inferred in jModelTest (https://github.com/ddarriba/jmodeltest2) as the best fit￾ting model [9]. In order to investigate the temporal signal, we regressed root-to-tip genetic distances from this ML tree against sample collection dates using TempEst v 1.5.1 (http://tree.bio.ed.ac.uk) [10]. The ML phylogeny was used as a starting tree for Bayesian time-scaled phylogenetic analysis using BEAST 1.10.4 (http://beast. community/index.html) [11]. We employed a stringent model selection analysis using both path-sampling (PS) and stepping stone (SS) procedures to estimate the most appropriate molecular clock model for the Bayesian phylogenetic analysis [12]. We tested a) the strict molecular clock model, which assumes a single rate across all phylogeny branches, and b) the more flexible uncorrelated relaxed molecular clock model with a lognormal rate distribution (UCLN) [13]. Both SS and PS estimators indicated the uncorrelated relaxed molecular clock (Bayes Factor = 4.3) as the best fitted model to the dataset under analysis. Besides, we have used the he HKY+G4 codon partitioned (CP)1 + 2,3 sub￾stitution model and the Bayesian Skyline coalescent model of population size and growth [14]. We com￾puted MCMC (Markov chain Monte Carlo) duplicate runs of 100 million states each, sampling every 10,000 steps. Convergence of MCMC chains was checked using Tracer v.1.7.1 [14]. Proper mixing of the MCMC was checked for ESS values >200 for each estimated para￾meter using Tracer 1.7. Systematic Biology. 2018;67(5):- 901–4). A Maximum Clade Credibility (MCC) trees was obtained from the tree posterior distribution using TreeAnnotator (http://beast.community/index.html) after 10% burn-in. Results Despite the short time since the beginning of the epi￾demic, the isolates analyzed have already exhibited a substantial degree of heterogeneity with differences in 15% of the sites, 11% of which were parsimony infor￾mative, thus indicating the presence of sufficient phylo￾genetic signal for further analysis, in agreement with the low level of phylogenetic noise shown by likelihood mapping (<7%). The root-to-tip vs. divergence plot of the full dataset showed high correlation between sam￾pling time and genetic distance to the root of the ML tree of the available sequences (R-squared 0.85), indicating substantial temporal signal and the possibility to calibrate a reliable molecular clock, despite the limited number of sequences and short sampling interval available. Bayesian model selection chose the Bayesian Skyline demographic model with an uncorrelated relaxed clock as the one that best fit the data. Molecular clock calibra￾tion estimated the evolutionary rate of the 2019-nCoV whole genome sequences at 6.58 × 10−3 substitutions site per year (95% HPD 5.2 × 10−3 – 8.1 × 10−3 ). Figure 1 A,B shows the MCC tree with Bayesian phy￾logeographic reconstruction of 2019-nCoV isolates. The probable origin of 2019-nCoV is, as expected, Wuhan with a state posterior probability (spp) of 0.93 dating back the time of the most recent common ancestor (MRCA) of the human outbreak to November 25, 2019 (95%HPD: September 28, 2019; December 21, 2019), while the MRCA of Bat SARS-like Coronavirus and related 2019-nCoV lineages dates back to February 22, 2011 (95%HPD: September 20, 2008; August 15, 2014) (Figure S1), which may suggest a relatively extended period of sub-epidemic circulation before the most recent events. The first evidence of 2019-nCoV dissemination appears to be, according to our phylogeographic reconstruction, from Wuhan, China, to Nonthamburi, Thailand, with an spp. of 0.96, followed by the emergence of two distinct lineages, one with further spreading in Nonthamburi, and the second one following a more complex pattern: from Nonthamburi to Zhejiang, Huangzhou (spp = 0.47), as well as from Zhejiang to Kanagawa, Kanto (spp = 0.62) and from Nonthamburi to Guandong, Zhuhai (spp = 0.45). The first reported US cases, in Chicago, Illinois and Seattle, Washington, appeared to be linked to Guandong, Zhuhai isolates, in agreement with reports of patients traveling back from that region of China before being diagnosed. Finally, our analysis identified the Bat SARS-like Coronavirus as the most probable origin of the 2019-nCoV (spp = 0.99). Discussion Very little is known about 2019-nCoV virus, including basic biology, animal source or any specific treatment. The substantial degree of genetic heterogeneity (15%) accumulated among human isolates during the past few months of the ongoing epidemic outbreak is not neces￾sarily surprising for an RNA virus that has been shown to be a measurably evolving population over short time spans [15]. However, our findings underscore the urgent need for further molecular surveillance and the develop￾ment of appropriate and an in-depth monitoring system capable of investigating viral mutation and transmission capabilities as 2019-nCoV unfortunately keeps spreading at a regional and potential global level. In other words, given the virus’s fast evolutionary rate and population dynamic, tracking the emergence of novel transmission routes and/or patterns should be considered a significant priority. 2 D. BENVENUTO ET AL