PCR-free libraries

[matth431]

Hi,

I think I've figured out a problem we have with a recent experiment. We have amplicons of ~400bp with a 5bp index added on to both the 5' and 3' ends (part of the PCR primers). The forward and reverse primers for each PCR have identical indices.

We then pooled equimolar amounts of 50 separate PCRs (each with its own unique 5' and 3' index) and ligated the standard TruSeq adaptors prior to library enrichment as per the TruSeq DNA protocol.

We get good sequence quality, but the paired reads do not match up in terms of index. For example, if read 1 sequences through index GATCA, we would expect read 2 to have the same index (or possibly reverse complement depending on which strand gets sequenced). However, we get a random assortment of read 2 indices.

I think the issue is chimera formation during the second round of PCR when amplifying with the Illumina primers. Because the mixed amplicons are fairly similar except for the indices, they are annealing so that the top strand might be 5'-GATCA index but bottom strand is 5'-TTCTT...

If I'm right, we could solve this using the new PCR-free DNA kits, but I'm uncertain as to DNA input and was wondering if anyone could help? The kit specifies input of 1-2ug of DNA, but we start from the A-tailing step (our amplicons are already at 400bp and blunt-ended so no need to fragment or end repair). My question would be, what sort of input DNA level would Illumina be expecting following the two bead cleanups in the fragmentation and end repair stages? Presumably there will be some loss of template prior to the A-tailing?

Thanks,

Matt

[mcnelson.phd]

Originally Posted by matth431

If I'm right, we could solve this using the new PCR-free DNA kits, but I'm uncertain as to DNA input and was wondering if anyone could help? The kit specifies input of 1-2ug of DNA, but we start from the A-tailing step (our amplicons are already at 400bp and blunt-ended so no need to fragment or end repair). My question would be, what sort of input DNA level would Illumina be expecting following the two bead cleanups in the fragmentation and end repair stages? Presumably there will be some loss of template prior to the A-tailing?

I've only ready through the protocol when seeing if the PCR-free kit would work for me, so I can't comment from a hands on perspective, but I have done the old standard TruSeq DNA kits. I believe that at the A-Tailing step Illumina expects that you'll have between 200-750ng of DNA, although the actual amount would be highly dependent on shearing and recovery yields, which you don't have to worry about it seems.

My suggestion would be to try 500ng as a happy medium if you go this route.

[matth431]

Originally Posted by mcnelson.phd

I've only ready through the protocol when seeing if the PCR-free kit would work for me, so I can't comment from a hands on perspective, but I have done the old standard TruSeq DNA kits. I believe that at the A-Tailing step Illumina expects that you'll have between 200-750ng of DNA, although the actual amount would be highly dependent on shearing and recovery yields, which you don't have to worry about it seems.

My suggestion would be to try 500ng as a happy medium if you go this route.

Thanks, that's interesting. We do end repair and A-tailing on size-selected libraries purified via the LabChip XT system, so based our input for this amplicon library on the recovery from that, which is only about 100-200ng total. We did get a lot of adaptor dimer, but were able to AMPure it out with a 0.6x cleanup.

The DNA-free protocol seems to use the same volumes of A-tailing mix and ligation mix/adaptor as the old kit. I may give it a go with the remaining DNA and see if it works.

Originally Posted by matth431

Hi,

I think I've figured out a problem we have with a recent experiment. We have amplicons of ~400bp with a 5bp index added on to both the 5' and 3' ends (part of the PCR primers). The forward and reverse primers for each PCR have identical indices.

If you are already working with amplicons, why don't you just design fusion primers? You'd get your amplicons with barcodes (on both ends if preferred) and adapter sequences in one go.

As much as I've heard from people with experience, ligating Illumina adapters to amplicons is inefficient.

[matth431]

Originally Posted by lorendarith

If you are already working with amplicons, why don't you just design fusion primers? You'd get your amplicons with barcodes (on both ends if preferred) and adapter sequences in one go.

As much as I've heard from people with experience, ligating Illumina adapters to amplicons is inefficient.

We won't be doing this routinely, and a set of fusion primers with 50+ barcodes would set us back ~£5000.