Question on multiplexing - long reads

[Drmadscientist]

Hi All

So I have 2-3 number of 2MB bacterial genome and two plasmids of 17 kb each length. I am looking at the feasilbility of multiplexing them in a single SMRT or MinIon flow cell. Has anyone ever multiplexed plasmids with genomic DNA?

Also, if you look at the data requirement, I would like say 1000x for both but then for PacBio they suggest you start with equal amounts of DNA for shearing. That will probably mean that I will get 1000x of bacterial genome and maybe 10000x of plasmid sequence which I do not want .

Same goes for MinIon multiplexing.

In Illumina you could do your adjustments for data requirement when you dilute your sequencing pools. Is the same possible here and has anyone tried?

Thanks
DMS

[nucacidhunter]

For PacBio you can use different quantity and pool based on molar concentration (after ligation step) to obtain relatively equal read coverage.

Points to consider:
1- adding twice the required coverage for plasmid as it might be supercoiled and resist shearing
2- depending on sheared DNA fragment size, it might be difficult to size select because plasmid DNA will shear to smaller fragments so you will need to shear gDNA to similar size which will exclude prepping 20kb library
3- There will be sequencing bias toward shorter fragments


I guess similar issues will apply for Nanopore as well

[SNPsaurus]

As nucacidhunter says, while there is worry that multi-copy plasmids will overwhelm a sample, in practice plasmids can be under-represented. In the below examples, the plasmids are part of the sample, not sequenced on their own, but they were most likely over-represented in the DNA used as input. We can see how they ended up in the sequencing library.
We multiplex samples in our PacBio runs, and call out contigs that are plasmids. For example, a recent run had a report that looks like (slightly redacted for privacy of the customer):

The assembly by Canu 1.7 was generated from 62280 reads, which yielded 601489362 bases for 150.37 read depth over the genome.
Read length histograms of reads used for assembly by Canu
-- Read length histogram (one '*' equals 51.25 reads):
-- 0 999 0
-- 1000 1999 229 ****
-- 2000 2999 92 *
-- 3000 3999 42
-- 4000 4999 36
-- 5000 5999 63 *
-- 6000 6999 76 *
-- 7000 7999 89 *
-- 8000 8999 129 **
-- 9000 9999 251 ****
-- 10000 10999 2320 *********************************************
-- 11000 11999 3588 **********************************************************************
-- 12000 12999 3206 **************************************************************
-- 13000 13999 2604 **************************************************
-- 14000 14999 2305 ********************************************
-- 15000 15999 1599 *******************************
-- 16000 16999 1127 *********************
-- 17000 17999 481 *********
-- 18000 18999 51
-- 19000 19999 2
-- 20000 20999 1
-- 21000 21999 0
-- 22000 22999 0
-- 23000 23999 0
-- 24000 24999 1

Contig: tig00000001|arrow with length: 4788128 has blast hit: E. coli complete genome
List of plasmids as identified by blast:
Contig: tig00000012|arrow with length: 143359 has blast hit: Escherichia coli plasmid DNA

Contigs identified as circular by Canu:
>tig00000012 len=143317 reads=314 suggestCircular=yes

So the plasmid had 314 reads for 143kb, around 20X read depth if assuming reads are 10kb.
The main genomic contig had 16,280 reads for around 35X read depth:
>tig00000001 len=4786778 reads=16280 35X read depth

Here's a second anecdote.
Contig: tig00000001|arrow with length: 4158188 has blast hit: Bacillus subtilis genome
List of plasmids as identified by blast:
Contig: tig00000018|arrow with length: 24327 has blast hit: Acidithiobacillus plasmid

Contigs identified as circular by Canu:
>tig00000001 len=4158151 reads=15701 suggestCircular=yes

Here's the contig header for the plasmid
>tig00000018 len=24327 reads=35

So in this example the main genome is a complete, circular assembly and uses 15k reads for 4Mb...also around 35X read depth. The plasmid is at 15X read depth.

Now, I looked for some example, any example, where the plasmid has a high read depth. This is the closest in our recent runs:
Contigs identified as circular by Canu (blank if none):
>tig00000001 len=2699848 reads=17396 suggestCircular=yes
>tig00000003 len=95082 reads=799 suggestCircular=yes

The main genome is at 65X read depth, and the plasmid is at 85X read depth. Higher, but not a worry about overtaking the representation in the sample.