Impact of meteorological parameters and population density on variants of SARS-CoV-2 and outcome of COVID-19 pandemic in Japan

Although vaccines have become available, emergence and rapid transmission of new variants have added new paradigm in the coronavirus disease-2019 (COVID-19) pandemic. Weather, population and host immunity have been detected as the regulatory elements of COVID-19. This study aims to investigate the effects of weather, population and host factors on the outcome of COVID-19 and mutation frequency in Japan. Data were collected during January 2020 to February 2021. About 92% isolates were form GR clades. Variants 501Y.V1 (53%) and 452R.V1 (24%) were most prevalent in Japan. The strongest correlation was detected between fatalities and population density (rs = 0.81) followed by total population (rs = 0.72). Relative humidity had the highest correlation (rs = −0.71) with the case fatality rate. Cluster mutations namely N501Y (45%), E484K (30%), N439K (16%), K417N (6%) and T478I (3%) at spike protein have increased during January to February 2021. Above 90% fatality was detected in patients aged >60 years. The ratio of male to female patients of COVID-19 was 1.35:1. This study will help to understand the seasonality of COVID-19 and impact of weather on the outcome which will add knowledge to reduce the health burden of COVID-19 by the international organisations and policy makers.

The first confirmed case of COVID-19 was reported from Wuhan, China in December 2019 [1]. As of 15 February 2021, about 110 million cases and 2.5 million fatalities of COVID-19 have been confirmed from more than 221 countries and territories globally [17,18].
Mutation at receptor binding sites (RBDs) of spike protein is mainly involved in the origin of recent variants of SARS-CoV-2 [29][30][31]. Deletion, indels and cluster of substitution point mutations at RBDs influence the origination of escape variants with a high growth rate. Numerous parameters of host and weather influence the mutational events [30][31][32]. Inside host body, the immune system, presence of coinfection and antiviral drug pressure are shaping mutations in coronavirus genome. Besides, parameters of weather specifically UV radiation, high ambient temperature, RH and snowfall have regulatory roles in evolving mutant viruses [22][23][24][25].
The principal objective of this study is to investigate the impact of regulatory factors specially, weather on mutation frequency in SARS-CoV-2 in Japan. Other objectives of this study are to specify the relationship between mutations at RBDs and host factors, and to determine the impact of weather on the outcome of COVID-19 pandemic in Japan. We also analysed the impact of population density, social gatherings and lockdown on the incidence and casualties in Japan. This study will create a better understanding of the effects of various regulatory factors on mutations of SARS-CoV-2 and outcome of COVID-19 in Japan.

Study regions and time period
This study determined the impact of weather on COVID-19 in Japan. Eight regions in Japan namely, Hokkaido (44°N to 143°E), Kyushu (32°5 ′ N to 130°8 ′ E) with 47 prefectures were included in this study. This study was conducted during February 2020 to February 2021.

Whole genome and mutational analysis of SARS-CoV-2 variants
The whole genome sequences of SARS-CoV-2 variants in Japan were analysed using Chromas 2.6.5 (Technelysium, Helensvale, Australia). Sequence homology of the selected whole genomes was deduced by using the nucleotide BLASTn program of NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Using the reference strains (NC_045512/Wuhan-Hu-1) and whole genome sequences of Japanese SARS-CoV-2 isolates multiple sequence alignment was performed by operating ClustalW Multiple Alignment algorithm in BioEdit 7.2.6 software. Clades were defined according to the presence of specific markers following the GISAID system [33]. Nucleotide and peptide sequences of whole genome of the selected isolates were analysed by using MEGA X to determine deletion, indels and substitution point mutations [9,22].

Statistical analysis
All data were sorted and analysed using unbiased statistical approaches. Appropriate correlation analysis was performed to determine the association between two variables by using a monotonic function of Spearman's rank correlation coefficient (r s ) [22]. The Spearman's rank correlation coefficients (r s ) were calculated between weather factors and cases as well as weather factors and casualties [22,25]. Furthermore, associations between host factors and mutation frequency, weather parameters and mutational events of SARS-CoV-2 were also determined by using the Spearman's rank correlation coefficient (r s ). The below equation is used to calculate the coefficient in this study: Here, 'n' represents the number of observations, 'd i ' represents the difference between the ranks of observations and r s represents the Spearman's correlation coefficient.

Temporal and spatial distribution of COVID-19 cases and fatalities
A sharp increase in both cases and fatalities of COVID-19 has been detected after October 2020 in Japan. About 75% of cases and fatalities have occurred during October 2020 to February 2021 ( Fig. 1a and b). Although the cases and casualties have increased, the case fatality rate has remained constant near 1.8 in Japan. All of the 47 prefectures in eight regions have been hit by the second wave of COVID-19. Kanto is burdened with the highest number of cases (215 000), whereas Kansai has the most number of fatalities (about 1800) ( Table 1). The number of cases in Kanto is followed by Kansai (∼77 000), Chubu (∼57 000), Kyushu (∼34 000) and Hokkaido (∼18 000), respectively (Table 1). Twenty prefectures burdened with the most number of cases and fatalities have been studied. The highest number of cases was detected in Tokyo (∼100 000) followed by Osaka (∼44 000) and other prefectures. The highest number of fatality was found in Osaka (∼1000), whereas Hokkaido had the highest case fatality rate ( Table 2).

Analysis of factors of weather
Amid parameters of weather, temperature, UV ray intensity, RH, rainfall, snowfall, wind speed, atmospheric pressure and sun hours were analysed for their role in COVID-19. Three temperatures were documented during January 2020 to February 2021. The mean of maximum temperature was about 15°C, and ranged from −3 to 41°C in eight regions in Japan. The mean of average temperature was 13°C and minimum temperature was 8°C in Japan during the last 14 months of COVID-19 pandemic

Epidemiology and Infection
( Fig. 2a-h). The mean values of minimum, maximum and average temperature varied ±6°C in eight regions. The lowest temperature was detected in Hokkaido (−16°C) in February 2020. A peak of three temperatures in eight regions was confined in August 2020 in Japan. Temperature started to decrease and both cases and fatalities of COVID-19 started to increase from September 2020. Sun hours or the amount of day light duration and UV ray intensity are important weather factors that regulate the spread and mutations of SARS-CoV-2. The lowest average UV intensity was 1 in Hokkaido during January 2021. The highest average UV intensity was 9 in Kansai during July 2020 (Fig. 3a). The average value of UV intensity ranged from 1 to 9 during the COVID-19 pandemic. One peak of average UV intensity was confined during April 2020 and another one during August 2020. After August 2020, UV index declined while the cases and fatalities increased. The sun hours ranged from 95 to 600 h per month. A peak of the sun hours was also detected in August 2020 (Fig. 3b).
The amount of rainfall and snowfall were also analysed in this study. The highest rainfall was detected in Kyushu (690 mm) in July 2020 followed by Chugoku (580 mm) and Tohoku (420 mm) (Fig. 4). Snowfall was not persistent in eight regions. Only seasonal snowfall was recorded in most of the regions in Japan. The highest rainfall was detected in Hokkaido during February 2020 and 2021.
RH, wind speed and atmospheric pressure are also significant factors of weather related to COVID-19. Average RH ranged from 45% to 90% during January 2020 to February 2021 in Japan ( Fig. 5a and b). The highest value of RH was recorded in Chugoku (90%) and Tohoku (90%) in July 2020, followed by Hokkaido (85%) in February 2020. A peak of average RH for eight regions in Japan was detected in July 2020. The highest average RH was detected in Hokkaido (80%) followed by Chugoku (78%) and Tohoku (77%), respectively. Atmospheric pressure varied around 1 atm during the pandemic in Japan.
Velocity of wind was recorded and analysed. The average wind speed per month was detected between 4 and 20 km/h in Japan. The highest average wind speed was detected in Kanto (16 km/h) followed by Kyushu (13 km/h) and Kansai (12 km/h), respectively ( Fig. 5a and b). The growth rate of cases in Kansai was the highest. The lowest average wind speed was detected in Tohoku during the COVID-19 pandemic.
The total population and population density of eight regions in japan were also analysed. Kanto (43.4 million) had the highest number of total population followed by Kansai (22.6 million), Chubu (21.4 million), Kyushu (14.5 million), Tohoku (8.9 million), Chugoku (7.3 million), Hokkaido (5.4 million) and

Correlation analysis between weather and COVID-19 pandemic in Japan
The impact of factors of weather on the outcome of COVID-19 was determined by Spearman's correlation analysis. Twelve factors of weather were considered on three time frames, namely on the day of the incidence, 7 days after and 14 days after the incidence in this study. Four parameters of COVID-19 namely, cases, fatalities, case fatality rate and growth rate were defined as the outcome of the pandemic. The Spearman's rank correlation was calculated between each factor of weather and outcome of COVID-19. Minimum temperature (min temp) on the day of the cases had the highest correlation with the cases (r s = −0.55) followed by average temperature (avg temp) on the day (r s = −0.51), sun hours on the day (r s = −0.49) and RH on the day (r s = −0.47), respectively. Both the total population (r s = 0.59) and population density (r s = 0.47) were strongly correlated with the increase of cases ( Table 3). The cumulative fatalities had the highest correlation with population density (r s = 0.81) followed by total population (r s = 0.72), RH (r s = −0.59) and minimum temperature (r s = −0.54) on the day, respectively. The highest correlation of case fatality rate was detected with RH (r s = −0.71) and sun hours (r s = −0.54) on the day. The growth rate of cases had also significant association with population density (r s = 0.51) and total population (r s = 0.43) ( Table 3) First, the correlation between maximum temperature and the number of total cases in each city was evaluated. Similarly, the correlation between minimum temperature and total cases, average temperature and total cases were determined. Besides, correlations among three temperatures with total fatalities in each city were also determined. The average temperature on the day of the cases had the highest correlation (r s = −0.675), followed by average temperature 7 days ago (r s = −0.547), maximum temperature on the day (r s = −0.512) and minimum temperature on the day (r s = −0.486). Maximum temperature on the day had the highest correlation with total fatalities (r s = −0.611). The correlation between temperature and COVID-19 cases and fatalities were negative, which indicated that at lower temperature the number of cases and fatalities increased. Among the factors of weather,  snowfall and atmospheric pressure were not significantly related to COVID-19 in Japan.

Distribution of frequency of mutations, age and genders in patients with COVID-19
Isolates containing various substitution points and deletion mutations were detected in Japan. About 16 000 whole genome from Japan were analysed for mutation frequency. Mutations were detected throughout the whole genome of isolates in Japan. Deletions at both the 5 ′ UTR and 3 ′ UTR were common. At spike protein, D614G was detected in about 100% isolates. Substitution mutation, T478I (55%) at spike protein were predominant during April 2020 to September 2020 ( Fig. 7a-d).
Deletion and substitution mutations at nucleocapsid (N) and other nsps were also frequent in Japanese isolates. Although substitution mutations, R203K, G204R at nucleocapsid and P323L at NSP12 were detected in ∼90% of the isolates in Japan, the frequency reduced during January to February 2021. Among 22 point mutations at N and other eight non-structural proteins, eight namely, N_S194L, N_R203K, N_G204R, NS3_Q57H, NSP2_T85I, NSP5_G15S, NSP6_L37F and NSP12_P323L were persistent throughout the COVID-19 pandemic in Japan (Fig. 7a-d).
Both the gender and age distribution of patients with COVID-19 in Japan were analysed. The ratio of male to female patients was 1.35:1 in Japan. However, female was predominant in the age group of 20-39 years and above 80 years ( Fig. 8a  and b). The highest frequency of fatality was detected in patients aged above 80 years (65%), followed by the age group 60-79 years (32%) (Fig. 8a and b).

Correlation analysis between mutation frequency and weather parameters
Novel coronavirus has acquired frequent mutations throughout the genome. The correlation between weather and mutation frequency as well as host factors and mutation frequency were determined for the isolates of Japan by Spearman's correlation analysis (Tables 4 and 5). Point mutation at RBDs namely, Y145del had strong correlation with maximum temperature (r s = 0.61), snow fall (r s = 0.41), N501Y with maximum temperature (r s = 0.46) and sun hours (r s = 0.49), N439K with minimum temperature (r s = 0.43), A222V with sun hours (r s = 0.51), E484K with sun hours (r s = 0.53), T478I with RH (r s = 0.49) and E780Q with RH (r s = 0.43). The highest correlation between mutations of nucleocapsid protein and weather was detected for N_I292T and UV (r s = 0.68), followed by snowfall (r s = 0.67) and rainfall (r s = 0.61). Significant correlation was detected between NSP5_G15S and UV (r s = 0.91), followed by NS8_S24L and  Table 5). Frequency of mutations and weather were positively related to each other. The impact of different host factors such as coinfection, age and gender variability on mutation frequency was also analysed. Significant correlation between host factors and frequency of mutation was detected (Tables 4 and 5).

Discussion
Evolution of new variants and increase in transmission rate of COVID-19 have infected millions of people within short time during the second wave globally [34,35]. Numerous factors including hosts health conditions, comorbidity, age, gender and weather parameters are regulating the transmission rate and outcome of the pandemic [36][37][38]. Parameters of weather always play significant roles in determining the outcome and severity of infectious viral diseases. The wave of COVID-19 pandemic is also shaped by meteorological parameters namely, temperature, RH, UV index, rain fall, snow fall, atmospheric pressure and wind speed [21][22][23][24][25][26][27][28][36][37][38][39]. Besides, direct man to man transmission is the major mode of spread of COVID-19. Gatherings of people during social, religious and sports events and exportation of new variants to specific country by immigrants are also regulating the spread of COVID-19 [21][22][23]. Previously different studies have found a significant impact of weather on coronaviruses infection [38][39][40]. High temperature and UV can reduce the survival period and viability of SARS-CoV-2 [21-24, 26, 27, 36-39]. Temperature above 40°C can reduce the viability of SARS-CoV, MERS-CoV and SARS-CoV-2 significantly under both laboratory and environmental conditions [40][41][42]. In this study, the correlation and impact of various regulatory factors on the spread and outcome of COVID-19 in Japan were determined. A sharp increase in cases and fatalities associated with COVID-19 was detected after October 2020 when the temperature and UV index started to decrease significantly. Various meteorological parameters are correlated with the COVID-19 pandemic [21,22]. This study detected a significant correlation of weather and outcome of COVID-19 in Japan, which is similar to previous findings in Bangladesh, China, USA, Turkey and Indonesia [21][22][23][24][25][26][27][28][36][37][38][39][40]. Three time frames were used to determine the correlations. We included more parameters and longer analysis for Japan compared to the previous studies. Except for atmospheric pressure, 11 other factors of weather were significantly correlated with the increase of cases, fatalities, growth rate and case fatality rate. COVID-19 incidence was strongly correlated with average temperature (r s = −0.51), minimum temperature (r s = −0.55), UV index (r s = −0.42), sun
hours (r s = −0.49), RH (r s = −0.47), total population (r s = −0.59) and population density (r s = −0.47). This statement is in good agreement with previous findings in Bangladesh and China [22]. Fatalities of COVID-19 was significantly correlated with population density (r s = −0.81), total population (r s = −0.72), RH (r s = −0.59) and minimum temperature (r s = −0.54). In a previous study, the association of temperature and COVID-19 fatalities was also reported from Bangladesh [22]. Furthermore, the case fatality rate and growth rate of cases were also affected by weather and population density in Japan. The findings on the impact of weather and population on COVID-19 case fatality rate and transmission rate are eccentric in this study. Sharif and Dey [22] and Şahin [36] detected a significant association of weather with COVID-19 cases in Bangladesh and Turkey, respectively. This study detected stronger correlation among weather and COVID-19 in Japan. Similar to previous studies, we also found significant correlation between UV index and COVID-19 cases and fatalities. Positive correlation was persistent between wind speed, population density, total population and COVID-19 cases and fatalities in Japan which has similarity to previous findings in Bangladesh, USA and China [22,25,39].
Among three time frames, the first day of incidence had the strongest correlation with the weather and total population and population density.
Other regulatory factors including duration of lockdown, availability of health facilities, availability of vaccine, tendency of the people to practice COVID-19-associated health rules, frequency of social and religious gatherings and regulation on immigrants have also shaped the wave of COVID-19 [22,39]. Before October 2020, COVID-19 was well controlled in Japan, but with the change in weather, increase of international immigrations and circulation of new variants both the cases and fatalities increased in Japan [20].
Various substitution point mutations throughout the gnome of SARS-CoV-2 accelerated the origin of new variants [29,32]. Among these point mutations, cluster mutations at receptor binding domains of SARS-CoV-2 are the most significant in determining the transmission capability and antibody neutralisation effect on the virus. Vaccines of COVID-19 have become available recently. However, emergence and spread of new variants have threatened the efficacy of vaccine globally. In Japan, GR (92%) was the most prevalent clade followed by GH (3%) clade. Since   14 Nadim Sharif et al.
December 2020, four variants with high significance including 501Y.V1, 501Y.V2, 452R.V1 and 484K.V2 have been detected in Japan. These variants have been involved in high transmission rate, escape from immune reaction and lower detection rate in several countries [29,32]. Diversity and frequency of mutations at spike protein have increased after October 2020 in Japan. About all of the isolates in Japan contained D614G at spike proteins, N_S194L and N_R203K at the nucleocapsid region. Variants with substitutions namely, K417N, T478I, N501Y, E484K, N439K and S477N at RBDs have been detected recently in Japan. Both deletion and substitution mutations at structural and non-structural proteins have become abundant in Japanese isolates. This study reported the cumulative mutation frequency in SARS-CoV-2 for the first time during the last 14 months in Japan.
We analysed the impact of weather, population and host factors on the origin and circulation of substitution point and deletion mutations in this study. A previous study in Bangladesh has reported significant correlation between average temperature and mutation frequency at ORF1ab and at S-D614G [22]. Several research studies on influenza virus have reported the impact of weather and UV radiation on the evolution of substitution mutations [41,42]. Similar to previous studies, we also detected notable correlation between weather and mutation frequency of Japanese isolates, and between host factors and mutations [21][22][23][24][25][26][27][28][36][37][38][39][40]. Weather factors including temperature, UV, RH, snow fall, rain fall and sun hours were significantly correlated with the frequency of mutations at receptor binding domains of spike proteins and other structural and non-structural proteins. Furthermore, age and gender of patients were also correlated with the frequency of important mutations. However, unlike previous studies in Bangladesh, we did not detect any significant correlation between comorbidity and mutation frequency in Japan [22].
In the demographic analysis, this study found that the male to female ratio of the infected patients was not significantly different. The outcome of COVID-19 was severe in patients aged above 60 years. Both death rate and hospitalisation rate were higher in patients aged above 60 years in Japan. The findings on both the distribution of age and gender of COVID-19 cases in Japan were similar to previous studies [19].
The impact of weather on the increase of cases and fatalities in Japan was evaluated in this study. This study has reported the association of weather and host factors with the circulation of isolates with mutations for the first time in Japan. A complete analysis on the present variants and clades of the circulating SARS-CoV-2 in Japan during the last 14 months was conducted in this study. This study will help to take proper implications by providing the crucial information on the impact of weather, population and host factors on the frequency of mutations and outcome of COVID-19. In future, studies including more data on mutation of isolates and on clinical outcomes can be conducted based on this research to create a complete scenario of the COVID-19 pandemic. This study will work as a significant resource of information for future studies on environmental correlation with the COVID-19 pandemic. Crucial information on circulating variants and mutations in this study will work as a guideline for future studies focusing on the evolutionary dynamics of COVID-19.

Conclusion
This is one of the early studies focusing on the impact of weather on the COVID-19 pandemic in Japan. This study has created a complete concept on the seasonal pattern of the cases and fatalities of COVID-19 in Japan during the last 14 months. This study included numerous factors that can affect the outcome of the pandemic to understand the dynamic of COVID-19. Significant correlation was detected between weather and cases, weather and fatalities, population and cases, population and fatalities, weather and mutations, host factors and mutations. Four recently evolved variants including 501Y.V1, 501Y.V2, 452R.V1 and 484K.V2 have been reported in Japan. This study concludes that not only preventive measures and interventions, but also weather has significant roles in shaping the outcome and severity of the COVID-19 pandemic. This study will help international health organisations and policy makers to understand the COVID-19 pandemic for taking appropriate steps to minimise the disease burden.
Financial support. This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.
Conflict of interest. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Data availability statement. Restrictions apply to the availability to the data that support the findings of this study.