Library generation

There is already a great online resource summarizing Illumina adaptors and sequencing chemistry (https://teichlab.github.io/scg_lib_structs/methods_html/Illumina.html) but I’ll summarize briefly here focusing on the Miseq i100 approach (which matches the NovaSeq 6000). I’ll also include some details on upstream steps (library generation & cluster generation).

The core idea is we need to add universal ends to our DNA of interest so that’s it’s “sequenceable” on Illumina’s instruments. In other words, we need to adapt our DNA. Along the way, we also make sure the DNA is within an acceptable size range (typically 200-800bp) for sequencing (a limitation set by Illumina’s technique) and add unique sequences (“indexes”) to the DNA to mark (“barcode”) separate samples so that we can pool and sequence a bunch of DNA from different sources together and then assign the results (“reads”) back to their original samples later. There’s a lot of jargon when it comes to sequencing but it’s all fairly sensible once you understand the fundamental steps.

There are countless ways to adapt your DNA of interest to include the indexes and constant end regions required for sequencing. How you choose is largely based on the material you’re starting with and the nucleic acids you’re trying to capture. In our lab, we do a lot of target site analysis where we sequence a specific region of interest in the human genome, usually after introducing reagents to modify that particular site in a population of cells. The results allow us to determine how efficiently our intended edits were introduced. I’ll make a separate protocol for target-site sequencing with all the details but for now what’s important is that we simultaneously amplify, index, and adapt our target DNA using back-to-back PCRs:

The final library structure

Details

• In PCR1, we are amplifying a site of interest (<1000bp in length) out of the genome using primers A1 and B1 which complement opposite strands of the target dsDNA and each contain extra, constant sequences on their 5’ ends which don’t complement the target site but are required for Illumina sequencing.

• In PCR2, we are further extending the constant regions on each end of the target DNA (shaded = shared sequence) while also barcoding the DNA in the sample by using primers A2 and B2 which contain short sequences that are unique to the individual sample. In other words, if we had 10 different samples we should need 10 different A2+B2 primer mixes each containing their own unique index sequences.

• Of course, in the end the DNA isn’t kinked like shown in the cartoon but a long, linear dsDNA product where we’ve added ~140bp of extra sequence in total (~70bp on each side). These added regions include unique sequences barcoding the sample as well as constant sequences to enable flow cell binding & sequencing.

• Why did we split this up into 2 separate PCR steps? Ideally, we’d doing everything in one step however Illumina’s sequences contain some palindromic regions across forward and reverse primers which often produce a lot of concatemers if you try to add everything in 1 step. Also, with two PCRs, we can kind of physically separate the process into (1) isolation of the target DNA and (2) final barcoding and adaptation.

Here, the region of interest (green) can be anywhere between 50-1000bp in length. Typically, we keep it <500bp so the full region can be covered during sequencing. Adapting the DNA adds another ~140 bp of sequence in total.

Technically, the only absolutely essential sequences that need to be added to our DNA of interest for compatibility with Illumina sequencing are the very 5’ ends of the dsDNA shown above (the “P5” and “P7” adaptors, highlighted in pink) as these enable binding to the instrument’s flow cell. However, we typically also add extra constant sequences designed by Illumina (“TruSeq” or “Nextera” adaptors, included as part of the PCR1 and PCR2 primers shown in yellow/orange) for ease in later steps.