HPRC sample information (human Pangenome)

Question

[turcoa1]

Hello all,

I am trying to download Nanopore seqeuncing samples from https://human-pangenomics.s3.amazonaws.com/index.html?prefix=working/HPRC_PLUS/ (human pangenome reference consortium). I have downloaded and unzipped HG03098_1.fast5.tar.gz and I cannot seem to understand how the files are broken down once unzipping. First of all, in HPRC there are 3 different fast5 gzipped files you can download. I do not know where to find information on what the difference is between the 3 gzipped files. Do I need to download all of these files, or does just downloading 1 gzipped fast5 give me all the info for the sample? I am also trying to basecall these reads and I am struggling to find out what flow cell was used (9.4.1 or 10.4.), and the chemistry type used to produce this information. The data is under the HPRC_PLUS directory and I have been struggling to find this information. If you are familiar with ONT reads from HPRC, please let me know if you have any info on this.

[GenoMax]

It does not look like that information is provided at this link. You may need to refer to the publication that is associated with this data.

[turcoa1]

https://human-pangenomics.s3.amazonaws.com/index.html?prefix=working/HPRC_PLUS/HG03492/raw_data/nanopore/

Here is the exact link to the fast5 files for a sample (HG03492). How's that?

[GenoMax]

I was referring to the information you were asking for. Type of FC/chemistry etc. I briefly looked at one HPRC publication and below was all I found. I am not sure which exact publication the data you linked is included in.

Base calling was performed using Guppy (v.4.0.11), with default parameters and the high-accuracy PromethION model (dna_r9.4.1_450bps_hac_prom.cfg).

On a different note, dorado is supposed to be able to auto-select a model though have not personally used that feature yet.

[turcoa1]

Yes I have not used that feature either. Thank you I did see this publication however I just wasnt sure if the sample I am utilizing was sequenced in the exact same way. Maybe I will go ahead and give the dorado automatic model selector a shot after converting the fast5's from HPRC to POD5s. On another note though, if I were to use the 10.4.1 model on 9.4.1 samples, do you think this would cause a lot of issues. Not sure if this would even work because I know the model you select is dependent on the flow cell and chemistry.

EDIT: the question is basically, can you use 10.4.1 basecall models on files produced using 9.4.1 - will this give better or worse results

[GenoMax]

the question is basically, can you use 10.4.1 basecall models on files produced using 9.4.1 - will this give better or worse results

My expertise with nanopore does not reach that far but I would not. There must be a reason ONT provides separate models.