MIRA 4 poor performance compared to 3.9.X

[Gorbenzer]

Dear all,
i'm having big issues using MIRA 4 for bacterial denovo sequencing: number of contigs generated by assembly with MIRA 4 is insanely higher than with MIRA 3.9.X.

Here are some comparison of number of contigs generated by assembly with the two MIRA versions:

MIRA 3.9.15 ---> MIRA 4

38 ---> 252
44 ---> 168
74 ---> 236
48 ---> 206


All these datasets are of good quality and give between 50X and 100X coverage.


Anyone knows the cause of this? Maybe some kind of default setting that was changed and now need to be manually set?

[maubp]

Sounds like a similar issue to this MIRA mailing list thread (which is probably the best place to ask about MIRA problems): http://www.freelists.org/post/mira_t...ct-to-MIRA-39X

[sklages]

Just curious, is the assembly really fragmented or are there just a few short contigs additionally created? Are the large contigs comparable in size and assembly quality?

[Gorbenzer]

It's me who started the post on MIRA mailing list.

We've done some testing and comparison with fully assembled genomes.

MIRA 4 gives more contigs but avoids assembly errors. We'll stick with MIRA since with 400bp single end reads it seems to work better and contigs number don't increase so much