Fig 1.
Initial symptom differences between cases from China and from the USA.
A) A Hasse Diagram depicting the likeliest path of discernible symptom onset, highlighting the most likely initial symptoms, for China (green) and USA (purple). B) Table of raw frequency data specifying the number and the percentage of patients that experienced fever and cough in the China and USA datasets. C) Abridged Hasse Diagram depicting the transition probabilities from no symptom to fever or cough in the most likely path of discernible symptoms for COVID-19 patients in China. The double arrow signifies that the transition probability of the most likely first symptom is more than double the transition probability second most likely first symptom. The error of these probabilities is 0.010. D) Abridged Hasse Diagram depicting the transition probabilities from no symptom to fever or cough in the most likely path of discernible symptoms for COVID-19 patients in the USA. The error of these probabilities is 0.063.
Fig 2.
The three likeliest paths of discernible symptom order in China and the USA.
A) A Hasse Diagram depicting the first (gold), second (silver), and third (bronze) likeliest paths of the order of discernible symptoms in China. In the case where the gold and silver lines converge, they both follow the same path. B) A Hasse Diagram depicting the first, second, and third likeliest paths of the order of discernible symptoms in USA. In the case where the gold and silver lines converge, they both follow the same path. C) The most likely path of discernible symptom order in China (green) and USA (purple). D) The second likeliest path of discernible symptom order in China and USA. E) The third likeliest path of discernible symptom order in China and USA. The errors of the probabilities of the paths using the datasets from China and USA dataset are 0.010 and 0.063, respectively.
Fig 3.
The likeliest paths of discernible symptom order in Hong Kong and Brazil.
A) The most likely paths of the order of discernible symptoms of COVID-19 patients in Hong Kong (blue) and Brazil (red). Dashed red and blue lines denote transitions where both regions’ most likely path is the same. B) Table of raw frequency data specifying the number and the percentage of patients that experienced fever and cough in the Hong Kong and Brazil datasets. (The study used for patients in Brazil reported the frequency of fever in 65,310 patients and the frequency of cough in 66,514 patients.) C) Abridged Hasse Diagram depicting the transition probabilities from no symptom to fever or cough in the most likely path of discernible symptoms for COVID-19 patients in Hong Kong. The double arrow signifies that the transition probability of the most likely first symptom is more than double the transition probability second most likely first symptom. The error of the transition probabilities is 0.017. D) Abridged Hasse Diagram depicting the transition probabilities from no symptom to fever or cough in the most likely path of discernible symptoms for COVID-19 patients in Brazil. The error of the transition probabilities is 0.019.
Fig 4.
The three likeliest paths of discernible symptom order in Japan before and after the D614G mutation became prominent.
A) The first (gold), second (silver), and third (bronze) likeliest paths of the order of discernible symptoms of COVID-19 patients before the D614G mutation became prominent. In the case where the gold and silver lines converge, they both follow the same path. B) The first, second, and third likeliest paths of the order of discernible symptoms of COVID-19 patients in Japan after the D614G mutation became prominent. C) Table of raw frequency data specifying the number and the percentage of patients that experienced fever and cough in the datasets of Japan before and after the D614G mutation became prominent. (The study used for symptom frequencies of COVID-19 patients after the outbreak of the D614G mutation reported symptoms of 2636 patients, except for cough, where only 2634 of the patients were recorded.) D) Abridged Hasse Diagram depicting the transition probabilities from no symptom to fever or cough before the D614G mutation became prominent. The double arrow signifies that the transition probability of the most likely first symptom is more than double the transition probability second most likely first symptom. The error of the transition probabilities is 0.002. E) Abridged Hasse Diagram depicting the transition probabilities from no symptom to fever or cough after the D614G mutation became prominent. The error of the transition probabilities is 0.036. F) A bar graph displaying the relative change, compared to the previous year, of search popularity of the terms, “cough,” “fever,” and “diarrhea,” in the months, January, February, March, and April 2020 calculated from Google Trends.
Fig 5.
The likeliest paths of discernible symptom order in cities with varying weather and in Japan with varying age.
A) The most likely path of discernible symptom order using the overall dataset from China is shown on the left. The middle and right columns are the most likely path of discernible symptom order of datasets in Shanghai, China and Osaka, Japan, when the Wuhan reference strain was prominent. The mean high temperature during the time that data was collected is written in red, the mean low temperature during the time that data was collected is written in blue, and the error of each implementation is written in purple. B) The most likely path of discernible symptom order using the overall dataset from USA is shown on the left. The other columns are the most likely path of discernible symptom order of datasets representing, from left to right, Detroit, New York, and Atlanta in the USA, when the D614G variant was prominent. The mean high temperature during the time that data was collected is written in red, the mean low temperature during the time that data was collected is written in blue, and the error of each implementation is written in purple. C) The most likely path of discernible symptom order for all patients in the Japan dataset (in the earlier time frame and characterized by the Wuhan reference strain) is shown on the left. All other columns are the most likely path of discernible symptom order by age groups from the overall set. The initial transition is written in green if it is consistent with the overall most likely path. The error of each implementation is written in purple.