Combining nanopore libraries

Question

[mrmrwinter]

Hi

I have two PromethION libraries, one standard, and the other with ultralong read chemistry.

To create a draft assembly i pooled the fastqs of both libraries and assembled from the result.

Now i am coming to nanopolishing the assembly and it requires the fast5 signal files of reads used in the assembly, which are still seperate and unpooled.

Can i just pool the fast5s? This seems difficult due to nanopolish needing the sequencing_summary.txt file, of which there is a separate one for each library.

Should i not have pooled the fastqs for assembly, and instead just use the ultralong reads to scaffold?

[GenoMax]

• multi_to_single_fast5 script on this page.
• poretools combine (LINK) - this may not exactly be what you need.

[samuel.a.odonnell]

Just to note, I think you were correct to have added the ultra-long reads for genome assembly, it will definitely help with contiguity

Also perhaps you want to try alternative polishing pipeline such as Racon and Medaka, which are much faster and appear to give similar results

Answer 1

[colindaven]

I think you can just cat the sequencing_summary.txt together if you want to pack the libraries together (not sure if they even have a header).

I had this exact issue with nanopolish. Also, I think you can add multiple separate libraries if my memory serves ... (but it was > 1 year ago).

To be specific, you can use the -d option many times:

nanopolish index -d fast5_1 -d fast5_2 x.fastq

[mrmrwinter]

Awesome, thanks. This is exactly what i needed