Considering hybrid assembly with metaspades or just using metaflye. How does metaspades handle overlapping long reads that do not overlap with small reads?

[jolespin]

hybrid assembly for metagenomes remains an experimental pipeline and optimal performance is not guaranteed.

If I have pretty deeply sequenced long reads and average depth illumina reads, would you recommend using metaspades with the --nanopore option or metaflye? I have deeply sequenced nanopore reads (7-9 GB gzipped) and Illumina reads (~2-3 GB gzipped).

For long reads that do not overlap with many short reads, are they assembled separately into contitgs/scaffolds or are they removed in spades/metaspades?

[asl]

SPAdes is a short-read assembler and uses long reads to supplement short read assembly, it does not assemble the. You can read the details in hybridSPAdes article: https://academic.oup.com/bioinformatics/article/32/7/1009/1743807

[jolespin]

Ok awesome, thank you. This is what I was expecting but just wanted to make sure.