tested on plants, animals ?

[colindaven]

Hi,

looks interesting. Is there any reason I should not try this on large 1+ GB eukaryotic genomes? Scalability, design issues etc?

Thanks
Colin

[tmaklin]

Hi Colin,

We've only tested and designed for bacteria, so the tool expects haploid sequences as input, and there might be some extra work required to postprocess the data if the input is from an eukaryotic genome. Otherwise I don't think there's any design issues that prevent using the algorithms on eukaryotes.

Building the SBWT index could have scaling issues in terms of memory usage, @jnalanko might be able to give you more info here. If you are using our CLI some of the subcommands don't support prebuilt indexes (we need to index the query to do variant calling and reference based mapping) so this may be an issue if the indexing is very slow or expensive.

If you want to try the scaling, there's a separate CLI for just building the indexes here: https://github.com/jnalanko/sbwt-rs-cli or you can use the build subcommand from kbo-cli.