This is great! Thanks for making, Marcel. I would like to try using a SS lib prep for the first time, so I would be very thankful for anyone’s SCR protocol pro-tips. I was thinking of using a uracil-tolerant SRSLY kit (https://www.claretbio.com/products/kits), which I think is the same approach that the SCR protocol uses. Does anyone have experience with this kit?
Thanks a lot for setting this up @Marcel Keller! I've also been using SCR and I've experienced some issues. Most notably, I cannot get rid of huge adaptor dimer peaks (see attached BA trace). I have started using a 1:1 ratio during the post-indexing purification (using SPRI beads) but even then the peak is still pretty big..
*Thread Reply:* That's a lot of dimers! Could you describe what you're working with - like sample type, input, and scr conditions.
*Thread Reply:* Thanks a lot for you reply and also for sharing the spotlight protocol! I tried to summarize some of the key steps below but I realize it's a lot and the sample type is very particular. Anyways, I am happy about any comments or tips!
To answer your questions: this is a library which was built using DNA extract from sedimentary ancient DNA, more specifically from an Arctic marine core. This particular sample had decent DNA concentrations (~4 ng/uL) as it is one of the youngest samples from the core. I have used an input of 14 ng and Tier 3 dilutions. However, as I go down in the core (older samples), DNA concentrations quickly become too low to be detected using standard Qubit measurements, so for those samples I have used Tier 5 dilutions and a 1:1 dilution of DNA extract and EB buffer as an input (so 10 uL of DNA extract and 10 uL EB). The older samples have an even higher dimer to library ratio. I am hesitant about using 20 uL of extract directly as I don't have that a lot of DNA extract for each sample (so if something went wrong I would need to re-extract).
In terms of indexing PCR: I have been using qPCR to determine the appropriate cycle number (based on the Gansauge et al. 2020 calculation) and I used two indexing replicates (20 uL volume) with an input of 6 uL pre-indexed library. After indexing, I pool the two replicates and continue with the SPRI bead cleanup.
*Thread Reply:* Thanks for the details! I think your post amp SPRI change to 1X was a reasonably way to reduce dimers (some have seen improvements with 2x rounds of 1.2X) but it's definitely difficult to significantly reduce dimers after amplifcation. Are you using a MinElute column or SPRI to purify after ligation?
If you're concerned with retaining extract then you may want to try spotlight. It's lower volume (10 µL extract max) compared to original and would be a better fit for your unquantifiable extracts. For the sample you've show, my recommendation would be to drop down another adapter tier - but at that point you're in spotlight range anyway so may want to dilute those extracts down for spotlight too.
*Thread Reply:* Thanks again! I will definitely try out spotlight. Regarding the post-ligation purification: I have been using MinElute columns.
*Thread Reply:* Would you expect to see lower adapter dimer formation when the indexing PCR is run in smaller volume replicates rather than one big volume? I'm seeing this trend but I haven't kept all other parameters constant..
*Thread Reply:* For the post ligation clean, the SCR protocol has an option for SPRI, which I recommend over MinElute. Spotlight also uses SPRI post ligation. In our hands, SPRI nearly always outperforms Minelute (increased yield, reduced dimers) while also being cheaper (when home-brewed) and higher throughput.
A couple things on indexing PCR. First, I think if you're seeing a trend for your samples then you should trust those observations. More generally, we don't see a difference in dimer percent when altering the input of indexing PCR if we control for cycle number and retain final indexing reaction conditions (50 µL indexing rxn, 1 µM each primers). But we primarily use spotlight and I don't process a ton of sediment.
There may be some sample specific things (inhibition) where reducing indexing input could lead to reduced dimer percent but I think it would be case-by-case.
*Thread Reply:* Alright, once again thank you very much! I will keep experimenting and if it turns out to be the samples, at least I have learned a lot in the process.
Hi all, I'm always happy to help troubleshoot the tech - here, over email, or zoom call. Two quick notes that may be generally informative:
We have a newly published variant of the tech (attached) that may be interesting to try if you often run the two lowest dilution tiers of the SCR. It's meant to be a replacement of those two tiers to deliver better conversion and lower dimers when working with <5-10ng of DNA. We typically work with extremely low to mid yield degraded samples so we near exclusively use this protocol.
Additionally, over the years we've ordered hundreds of SCR oligos and have found HPLC purification is a massive contamination risk. We recommend standard desalting the adapters and splints, rather than the original recommendation of HPLC.
*Thread Reply:* Hey Pete, they should be very similar (~10mM final). There could definitely be a typo somewhere.
Spotlight linked here is 10 mM. The supplement of the SCR protocol calls for 3.75 µL of SCR buffer (containing 132 mM MgCl2) per rxn, over a 50 µL rxn is 9.9 mM
*Thread Reply:* You are right — the typo was in my calculations… oops. 😬 (missed the 1 in 132…) Thanks for your help, Josh!
*Thread Reply:* I didn’t have any PCR failure and indeed I was using qPCR and predicting the number of cycles I needed for library amplification. Nevertheless, the biggest problem was adapter dimers for the lowest tier adapter dilutions which in turns depend on the DNA concentration. I did have leftover ligated non-amp library simply to have back up for later amplifications and to add complexity to deeply sequenced libraries for the same sample (which worked quite well). Have you tried cleaning the libraries from inhibitors with Zimo kit perhaps?