To rebuild the database, we need to first download available versions of the SARS-Cov-2 genome. Visit https://www.ncbi.nlm.nih.gov/labs/virus/vssi and click Search by virus. In the search box type SARS-Cov-2 and click on taxid:2697049. A new page will appear with the list of genome sequences. On the left panel, locate the tab Nucleotide Completeness and check the box for complete. The list of genome sequences should automatically update, keeping on those genomes that are complete. On the top-right, click the button Download. You will need to download two files. First, under Sequence data (FASTA Format) download Nucleotide using the default options that appear. Then, under Current table view result download CSV format using the default options that appear. These downloads will result in the following two files.
sequences.csv
sequences.fasta
To build the database of mutations, run the following command in the terminal. The script may take approximately an entire day to complete.
python3 mutations_parallel.py # This script does the same computations as mutations.py distributed across multiple CPU cores
The script saves the results in a file called mutations.csv
The file mutations.csv list the point mutations observed in the SARS-Cov-2 genome relative to the reference genome NC_045512. Each point mutation is represented as a symbol, a number, and another symbol. The first symbol represents the original nucleotide according to the reference genome. The number represents the position of the mutation in the reference genome. The last symbol represents the nucleotide after the mutation. Each mutation is listed with the earliest date it is observed along with the accession code indicating the genome where the mutation first occurred.
There are two limitations with this analysis. The first limitation is that our analysis excludes insertions and deletions. This is because a preliminary analysis revealed that insertions and deletions occur almost exclusively at the ends of the genome, making it unclear if these insertions and deletions are sequence artifacts or actual mutations. The other limitation is that there is no way to determine if the remaining point mutations are actual mutations or sequencing error. Keep these limitations in mind going forward.