POD5 streaming data functionality

[harisankarsadasivan]

Hello,

Could you please explain the streaming functionality in C/C++ if I were to extract raw data for Read Until/selective sequencing?
I know a FAST5 has to be completely written before it can be read. Does POD5 have any advantages for Read Until?
How would future chunks of a read be handled? Appended to the same file or new file?

[0x55555555]

Hello,

Could you please explain the streaming functionality in C/C++ if I were to extract raw data for Read Until/selective sequencing?

POD5 is an on disk file format for nanopore read data - its two table approach supports writing partial read data to disk, and then appending to and completing reads later. You could extract the data for downstream analysis, but I would not expect you to do this in an application that feeds back into the sequencing process, like adaptive sampling does.

I know a FAST5 has to be completely written before it can be read. Does POD5 have any advantages for Read Until?

I don't expect this to have any direct benefits to read until, as this requires much lower latency access to small amounts read data - this is why we provide a specific API for read until.

How would future chunks of a read be handled? Appended to the same file or new file?

One read is always contained within one file - It would be technically possible to spread data across multiple files, but I dont feel like this would be useful to readers. Right now conceptually signal data without a record in the read table is considered "orphaned", and would not be read by the tools we provide.

Hope that helps,

• George

[iiSeymour]

You might also be interested in the docs and C++ examples here @harisankarsadasivan