hyeshik / poreplex

A versatile sequenced read processor for nanopore direct RNA sequencing
Other
78 stars 13 forks source link

Poreplex does not detect basecalled fast5 files #10

Closed SaraJeanne08 closed 5 years ago

SaraJeanne08 commented 5 years ago

Hi I am running the following on my MiniKNOW (albacore) fast5 files from a direct RNA seq run but Poreplex errors out: poreplex -i . -o /prod/Nanopore_Seq_Data/20181001_Sept17_18_2018_Trimmed_Fastq/ --trim-adapter --keep-unsplit -p 38

It states no basecalled files are found... but these are basecalled fast5 files - direct output from my MinION run. Here is the log file:

2018-10-01 13:51:53,768 Starting poreplex version 0.1 2018-10-01 13:51:53,768 Command line: /home/sdunaj/.local/bin/poreplex -i . -o /prod/Nanopore_Seq_Data/20181001_Sept17_18_2018_Trimmed_Fastq/ --trim-adapter --keep-unsplit -p 38 --b asecall 2018-10-01 13:51:53,768 == Analysis settings ====================================== 2018-10-01 13:51:53,768 Input: . 2018-10-01 13:51:53,769 Output: /prod/Nanopore_Seq_Data/20181001_Sept17_18_2018_Trimmed_Fastq/ 2018-10-01 13:51:53,769 Processes: 38 2018-10-01 13:51:53,769 Presets: rna-r941.cfg 2018-10-01 13:51:53,769 Basecall on-the-fly: Yes (albacore 2.3.3) 2018-10-01 13:51:53,769 Trim 3' adapter: Yes 2018-10-01 13:51:53,769 Filter concatenated read: No 2018-10-01 13:51:53,769 Separate by barcode: No 2018-10-01 13:51:53,769 Real-time alignment: No 2018-10-01 13:51:53,769 FASTQ in output: Yes 2018-10-01 13:51:53,769 FAST5 in output: No 2018-10-01 13:51:53,769 Basecall table in output: No 2018-10-01 13:51:53,769 =========================================================== 2018-10-01 13:51:53,769 2018-10-01 13:53:55,750 Finished.

Thank you!

hyeshik commented 5 years ago

Hi @SaraJeanne08,

MinKNOW and albacore produce different outputs for the basecalls. In fact, the live basecaller included in MinKNOW and albacore are the different ones. By far, poreplex is tested only with the FAST5 basecalls from albacore. If you'd like to feed the FAST5 files directly from MinKNOW to poreplex, you'll need to add --basecall option to the poreplex command line to let poreplex use albacore to re-basecall the signals. The current version of poreplex, unfortunately, does not play well with the most recent version of albacore, 2.3.3. Please install albacore 2.3.0 instead or apply a patch for the incompatibility issue for this time.