At the risk of starting a flame war, lvs is mostly correct.
These are NOT whole gene deletions, they are nucleotide deletions. Genes are usually longer than 33 nucleotides, by my count the largest deletion he sees. For reference, that’s 11 amino acids, rule of thumb for an average protein length is about 200 amino acids.
The author highlights that many of these deletions are not codon aligned, which means that if you make this mutation, the resulting protein is highly unlikely to be translated properly. This is true, however devoid of context.
Most of the mutations are very rare. The highest frequency one occurs 400 or so out of at minimum 10,000 examples. Next one is 36/10,000. It’s possible these matter, it’s also possible these viral sequences are just low quality and the mutations are an artifact. It’s also possible that these mutations don’t actually matter to viral function for whatever reason. For example, if the mutation is at the end of a protein, it could screw up everything downstream but the protein before the mutation works just fine, i.e. mutation doesn’t matter. It’s just hard to tell with the data in this article.
The concept that maybe there are attenuated strains isn’t completely bogus, there is some thought a Singapore strain (with a massive deletion of 380 nucleotides, or 10x what we see here [1]) might be less infective. This is speculation and we don’t know yet, but that sounds maybe possible to me as a PhD Biochemist (if we’re gonna get all credential-spreading here).
However saying “The really interesting deletion mutations are those in the non-structural accessory genes. These are the genes that are likely to play an important role in pathogenicity and it would be expected that some of these mutations may make the virus less dangerous (attenuated).” Is really pushing it. I don’t know of any science to believe this one way or another, viruses are pathogenic for a variety of reasons, both accessory and core (whatever that means) proteins matter.
In short, given the uncertainty, we can hope there’s a less infectious strain out there, but this article provides precious little evidence to that effect.
1.) https://www.biorxiv.org/content/10.1101/2020.03.11.987222v1....