Short mixed amplicon libraries

taylor_wilcox · 07-05-2017, 01:21 PM

I'm interesting in sequencing libraries composed of multiplexed PCR products with amplicon lengths running from 50 - 150 bp. A colleague expressed concern about sequencing bias for shorter amplicons on the Illumina platforms (likely will want to use MiSeq).

Does anyone have experience sequencing libraries of short, mixed-length amplicons/insert sizes? If so, can you comment on the level of sequencing bias? If sequencing bias was a problem, do have a good working solution (e.g., pooling amplicons of different lengths at different relative molarity)?

GenoMax · 07-05-2017, 07:56 PM

What kind of sequencing bias are you referring to? If you are going to be mixing products of different length then even the low nucleotide diversity (which may be an issue with amplicon sequencing) would not be a problem.

taylor_wilcox · 07-05-2017, 08:21 PM

Sorry I didn't make that clear. I mean a bias towards smaller amplicons (~50 bp) relative to larger amplicons (~ 150 bp). For example, would I expect a large drop in coverage for the 150 bp amplicons relative to the 50 bp amplicons when pooled equimolarly on the same lane?

GenoMax · 07-05-2017, 08:25 PM

It is true that shorter inserts tend to cluster more efficiently on the flowcell than longer ones. You could use less amounts of shorter ones and try to bias a uniform representation.

Someone with more practical bench experience may have more specific tips for you now we know what you are working with.

thermophile · 07-06-2017, 11:09 AM

length bias still exists on illumina flowcells but it's not anywhere close to as strong as it was with 454. You may need to tweak your load ratios a small amount but not much. 5-10% as a complete guess

fanli · 07-10-2017, 11:09 AM

I'm curious about this too - we've held back from running different amplicons together precisely because it'd be a complete shot in the dark as to how much to tweak loading concentrations.

Here is some data from one of our V1V2 runs (as these will have some naturally occurring length variation). No clear relationship between amplicon size and number of reads output. Note that amplicon size refers to the average size of the largest peak and all libraries were pooled equimolar for sequencing. These are from the Tapestation, fwiw.

thermophile · 07-11-2017, 06:47 AM

My primary type is 16s v4 (~300bp). I've mixed up to 30% of a run with ITS2 (~250-400bp) or custom ~150bp amplicons with no issues and no noticeable decrease in reads for the longer amplicons on the run but I'm looking for a minimum per sample rather than average

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Problem with TruSeq Custom Amplicon Libraries	HelenaSC	Sample Prep / Library Generation	2	07-02-2017 10:52 AM
Differential Expression Using mixed libraries (truseq, nextera)?	feralBiologist	Bioinformatics	3	11-13-2013 05:56 PM
duplicate reads removal for amplicon libraries?	ngseq	Illumina/Solexa	0	12-10-2012 01:50 PM
detecting deletion in amplicon sequencing of mixed population	Josim	Genomic Resequencing	0	11-10-2011 06:24 AM
PubMed: Methods for generating shotgun and mixed shotgun/paired-end libraries for the	Newsbot!	Literature Watch	0	04-11-2009 06:40 AM

07-05-2017, 01:21 PM	#1
taylor_wilcox Junior Member Location: USA Join Date: Oct 2016 Posts: 3	Short mixed amplicon libraries I'm interesting in sequencing libraries composed of multiplexed PCR products with amplicon lengths running from 50 - 150 bp. A colleague expressed concern about sequencing bias for shorter amplicons on the Illumina platforms (likely will want to use MiSeq). Does anyone have experience sequencing libraries of short, mixed-length amplicons/insert sizes? If so, can you comment on the level of sequencing bias? If sequencing bias was a problem, do have a good working solution (e.g., pooling amplicons of different lengths at different relative molarity)?

07-06-2017, 11:09 AM	#5
thermophile Senior Member Location: CT Join Date: Apr 2015 Posts: 243	length bias still exists on illumina flowcells but it's not anywhere close to as strong as it was with 454. You may need to tweak your load ratios a small amount but not much. 5-10% as a complete guess __________________ Microbial ecologist, running a sequencing core. I have lots of strong opinions on how to survey communities, pretty sure some are even correct.

07-11-2017, 06:47 AM	#7
thermophile Senior Member Location: CT Join Date: Apr 2015 Posts: 243	My primary type is 16s v4 (~300bp). I've mixed up to 30% of a run with ITS2 (~250-400bp) or custom ~150bp amplicons with no issues and no noticeable decrease in reads for the longer amplicons on the run but I'm looking for a minimum per sample rather than average __________________ Microbial ecologist, running a sequencing core. I have lots of strong opinions on how to survey communities, pretty sure some are even correct.

07-05-2017, 07:56 PM	#2
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,080	What kind of sequencing bias are you referring to? If you are going to be mixing products of different length then even the low nucleotide diversity (which may be an issue with amplicon sequencing) would not be a problem.

07-05-2017, 08:21 PM	#3
taylor_wilcox Junior Member Location: USA Join Date: Oct 2016 Posts: 3	Sorry I didn't make that clear. I mean a bias towards smaller amplicons (~50 bp) relative to larger amplicons (~ 150 bp). For example, would I expect a large drop in coverage for the 150 bp amplicons relative to the 50 bp amplicons when pooled equimolarly on the same lane?

07-05-2017, 08:25 PM	#4
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,080	It is true that shorter inserts tend to cluster more efficiently on the flowcell than longer ones. You could use less amounts of shorter ones and try to bias a uniform representation. Someone with more practical bench experience may have more specific tips for you now we know what you are working with.