PacBio vs. MiSeq

Hello all,

Hope you guys are doing well under pandemic.

I have a question about the sequencing platform.

As you know, PacBio provides more resolution of taxonomy profiling like species level while MiSeq platform generates only family or genus level taxonomy. Also, MiSeq often showed the inflation of the number of taxa present. However, some researchers worried about the error rate of PacBio platform. I’m also worried about low total number of reads per run in case of PacBio.

Given this condition, which one is generally suggested for normal microbiome analysis?

The error rate in long-read sequencing platforms has dramatically dropped in recent years. Per-based error rate of PacBio CCS reads (<0.1%) is even lower than that of Illumina short reads (0.1-0.5%). There are several publications on sequencing full-length 16S rRNA gene using highly accurate long read sequencing (PacBio, Nanopore, synthetic long read).

It’s stated in the PacBio website that their new protocol delivers more than 30,000 reads per barcode at 96-plex in a single Sequel II System run. If that’s not enough for your samples, you can consider suqucencing less samples in a run.

That depends on the question you want to answer. Budget allowing, you can consider highly accurate long read sequencing such as PacBio CSS. It gives you higher taxonomic resolution and more reliable phylogeny. Otherwise, standard short-read sequencing should be just fine. For more discussions, check this recent talk by Dr. Benjamin Callahan on “amplicon sequencing in the era of highly-accurate long reads”.


Thank you for your kind answer.
I’m trying to analyze a bacterial community of endophytes where a few microbes such as Pseudomonas and Bacillus sp. are commonly dominant. That’s why I wanna more resolution of taxonomy profiling. The data generated from MiSeq platform using V3-V4 region is not enough to figure out each ASV can represent one specific species.
I’m gonna consider PacBio CSS if we have enough money.

Hi @hyuck1857,

Is this because the specific species are all the same over that region or because the database you’re classifying against doesn’t have species level resolution? I’m not trying to turn you off either method, I’m just curious what the parameters of the problem are, exactly. Because if it’s a biological resolution issue, then a longer sequence will help, but if it’s a different issue, a longer sequence may not.