The Albertsen lab has recently put out a competition/challenge for read error correction

I only found out about this today, and I think the conference where high-ranking participants were going to be mentioned has just finished, but the data is all public and there's no reason why this can't be continued in the future as a benchmarking test for nanopore base calling and/or read error correction.

Data: Nanopore reads (as called FAST5 files) can be found here. The initial called FASTQ files can be found here. Reference sequences (from which the reads were generated) can be found here.

Every [passed] read should be a 2D nanopore read with the following sequence structure, where the decamer, NNNNNNNNNN, is the unique molecular identifier and is attached to common primer sites at the read extremities:

<—- fragment of SSU cDNA molecule—->

So... what's the best consensus accuracy you can get from these reads? How did you get that accuracy?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.