This webinar focused on how the JGI employs single-molecule, long-read DNA sequences to aid with genome assembly and transcriptome analysis of microbial, fungal, and plant research projects. More recently, these long-reads have been applied to DNA methylation analysis as well as improving metagenome assemblies.
Q&A from the Webinar:
Input & sequencing
Q(uestion): How much DNA is needed for “regular” vs the ultra-low input library creation protocols for PacBio?
A(nswer): Regular PacBio typically requires ~ 1.5-10 ㎍. Low-input protocols are continuously improved, but currently require input in the ~ 30 – 100ng range. Note that beyond quantity, input quality (i.e., presence of “intact” high-molecular weight DNA) is an important factor as well for PacBio metagenomes.
Q:How many PacBio metagenomes can I submit? How deeply are metagenomes sequenced?
A: For the next call (CSP New Investigator), up to 4 long-reads (PacBio) metagenomes can be included, which do not count toward the 1Tb cap. PacBio metagenomes are currently sequenced without pooling, which yields ~ 180 Gb of data per library.
Q:How much DNA/coverage is required for DNA modification detection?
A: DNA modification detection can only be performed from “regular” PacBio, since the low-input protocol involves PCR amplification of the library. 6mA and 4mC (~50x depth), 5mC (~250x depth required). Methylation detection currently only works using subread data, PacBio is working on support to use CCS data as input.
Data analysis
Q: What types of analysis support does JGI provide for DNA modification detection?
A: DNA modification detection can be performed on metagenomes as a custom analysis, i.e. is not part of the standard analysis pipeline but can be done on request by the PI. We currently use the ‘Base Modification Analysis’ tool within PacBio SMRT link, which works well for 6mA and 4mC (~50x depth), and can detect 5mC (~250x depth required).
Q: Can DNA modifications be detected in metagenomes?
A: Yes, DNA modifications can be searched for and identified in metagenomes, although not all types of DNA modifications are being detected currently.
Q: Does JGI support hybrid long-read / short-read metagenome assemblies?
A: At this point, we have not seen substantial improvement with hybrid assemblies when compared to long-reads-only assemblies, especially with CCS / Hi-Fi PacBio reads (i.e., reads with very low error rate), so our pipeline performs long-reads-only assemblies and not hybrid.
Q: Can short-read metagenomes be mapped to long-read metagenome assemblies by JGI and uploaded to IMG?
A: Depending on the project setup, we can include as part of our data analysis plan mapping of short-reads datasets (metagenome or metatranscriptome) to a relevant long-reads metagenome assembly. Integration of this type of “external mapping” into IMG is currently being revamped but we can already provide raw mapping files as well as gene- and contig-level coverage.
Other related products
Q: Does JGI offer long-read metatranscriptome? (for e.g. full 16S, RNA virus, etc)
A: Long-read metatranscriptome / RNA metagenomes are on our radar, but we don’t offer these at this time.
Q: How does the correction of the extended PacBio compare to doing polishing with Illumina reads?
A: The JGI has found that it’s not necessary to polish with both, and generally it’s more efficient just to make a single library tape. If you do have both Illumina and PacBio, the recommendation there is to assemble and error-correct with the PacBio data only, and then do a subsequent polishing round with the Illumina data. It’s not strictly necessary, as you can get finished quality microbial data polishing with just the PacBio data.