Help with MinION sequencing data species identification

Question

Hi I'm new to bioinformatics and have just completed my first run on the MinION (long read sequencing Oxford Nanopore Technologies). I was hoping someone could direct me towards R packages, workflow, tutorials or guides that will help me identify species that are present in my sample mainly for fungi as my amplicons had ITS1 and ITS2 regions.

I have been watching Youtube tutorials on RStudio for over a week and I feel like I'm no closer than I was when I started. I have both fastq and fast5 files and I tried to run the fastq files using EPI2ME, however, the database is lacking in fungi. I have tried using NCBI's BLAST but I can only run <5kb reads and most of my sequencing reads are >6kb. Also, I've tried using Galaxy to convert my fastq files to fasta but I'm not sure how to run my long reads for species identification using Galaxy or if that is even possible.

Any help would be greatly appreciated. Thanks.

Answer 1

There are a number of useful metagenomics tools out there, including Centrifuge and Kraken2, as @gringer has mentioned.

Also, FWIW, the WIMP tool from ONT, a Centrifuge based metageonomic classifier, does include fungus sequences derived from RefSeq, though it may not include the species you are looking for.

If you want to use NCBI BLAST, you could split your long reads into pieces that are <5kb (there are a number of tools out there for FASTQ manipulation - my current favorite tool for this is seqkit, where you could do something like this):

seqkit sliding -s 4900 -W 4900 test1.fastq

The above will split your FASTQ reads into chunks of 4.9kb. You can adjust the -s parameter if you instead want overlapping chunks.

From a GUI standpoint, you can use Galaxy for the above by making a workflow with the following steps:

1. Trim Sequences (enter reasonable numbers here, to trim each of your reads to <5kb)
2. FASTQ to FASTA converter
3. NCBI BLAST+

Another possibility is QIIME2, which can be used with the UNITE fungal training set. See recent reference (looking at the ITS1/ITS2 regions), UNITE, and QIIME2.

Long term, you may want to look into using the command line, as it is much more powerful and customizable, but if BLAST gives you reasonable results for your reads, the above process in Galaxy may work for you.

Answer 2

I've found Centrifuge and Kraken2 to be useful for species classification. There are pre-prepared databases that include Fungi, e.g. see here for the Kraken 2 "PF" database, which is the Standard database plus protozoa & fungi.

Can't really provide much more help than that with more information about what you've done.

You'll need a computer with around 64 GB of memory in order to be able to run these programs, and it's probably not the best idea to jump right into command line compilations when starting out... but it looks like you're up for a challenge given that your first MinION run has involved jumping into the deep end with a fungi metagenomic project.