Background COVID-19, caused by the novel SARS-CoV-2, is considered the most threatening respiratory infection in the world, with over 40 million people infected and over 0.934 million related deaths reported worldwide. It is speculated that epidemiological and clinical features of COVID-19 may differ across countries or continents. Genomic comparison of 48,635 SARS-CoV-2 genomes has shown that the average number of mutations per sample was 7.23, and most SARS-CoV-2 strains belong to one of 3 clades characterized by geographic and genomic specificity: Europe, Asia, and North America. Objective The aim of this study was to compare the genomes of SARS-CoV-2 strains isolated from Italy, Sweden, and Congo, that is, 3 different countries in the same meridian (longitude) but with different climate conditions, and from Brazil (as an outgroup country), to analyze similarities or differences in patterns of possible evolutionary pressure signatures in their genomes. Methods We obtained data from the Global Initiative on Sharing All Influenza Data repository by sampling all genomes available on that date. Using HyPhy, we achieved the recombination analysis by genetic algorithm recombination detection method, trimming, removal of the stop codons, and phylogenetic tree and mixed effects model of evolution analyses. We also performed secondary structure prediction analysis for both sequences (mutated and wild-type) and “disorder” and “transmembrane” analyses of the protein. We analyzed both protein structures with an ab initio approach to predict their ontologies and 3D structures. Results Evolutionary analysis revealed that codon 9628 is under episodic selective pressure for all SARS-CoV-2 strains isolated from the 4 countries, suggesting it is a key site for virus evolution. Codon 9628 encodes the P0DTD3 (Y14_SARS2) uncharacterized protein 14. Further investigation showed that the codon mutation was responsible for helical modification in the secondary structure. The codon was positioned in the more ordered region of the gene (41-59) and near to the area acting as the transmembrane (54-67), suggesting its involvement in the attachment phase of the virus. The predicted protein structures of both wild-type and mutated P0DTD3 confirmed the importance of the codon to define the protein structure. Moreover, ontological analysis of the protein emphasized that the mutation enhances the binding probability. Conclusions Our results suggest that RNA secondary structure may be affected and, consequently, the protein product changes T (threonine) to G (glycine) in position 50 of the protein. This position is located close to the predicted transmembrane region. Mutation analysis revealed that the change from G (glycine) to D (aspartic acid) may confer a new function to the protein—binding activity, which in turn may be responsible for attaching the virus to human eukaryotic cells. These findings can help design in vitro experiments and possibly facilitate a vaccine design and successful antiviral strategies.