Web Bio Tools is an experimental collection of browser-based
bioinformatics utilities written in Rust and compiled to
WebAssembly. Right now, it supports pairwise sequence alignment.
Was this project vibe coded?
Pretty much. In fact, this FAQ is the exception as it was mostly
hand written. The other big manual contribution was choosing the
example sequences.
Where are the example sequences from?
All of them are examples of alignments to the Comprehensive Antibiotic
Resistance Database (CARD, a database of antibiotic
resistance genes).
The very similar sequence is from our dog gut microbiome catalog
(which is not fully public yet).
The very dissimilar sequence is referenced in the argNorm
paper as the two ErmA genes, which were reported at the same time
(in 1985) and thus got the same name but are not closely related.
If the algorithms were vibe coded, how can we trust any of it?
They were compared to the implementations in BioPython and found to
be equivalent. These tests use the
[hypothesis](https://hypothesis.readthedocs.io/) library to check
randomly-generated examples.
The tests were also vibe coded.
When I save a result do I have to log back in to see it?
No, there is no logging in, everything is running on your browser.
Saved results are local to you (they are saved in your browser). If
you share them, all the information is present in the the
URLs.