When there is a lot of contigs, makeFurDb is very slow

[wangzhichao1990]

Hi,

When there are a lot of contigs, makeFurDb is very slow. The following figure shows the statistical results of the neighbor genomes. Is there a way to increase speed? I am using the latest docker version. Thanks.

[haubold]

To a first approximation, each neighbor sequence is turned into a suffix array. Since the computation of a suffix array comes with a performance overhead, the analysis of very many sequences in the neighborhood will slow down makeFurDb. One way to speet things up is to concatenate sequences into fewer, longer chunks. In the limit of concatenating all neighbors into one sequence, memory consumption is maximal and might outstrip the avable RAM. Hope this helps.