Caitlin M Stewart, Jahan-Yar Parsa, Jeremiah Li - Mar 26, 2024

Low-pass whole genome sequencing from canine saliva on FTA cards: A story of samples to solutions

Sequencing projects come in many different forms. At Gencove, we’ve seen and supported a broad range of species and sample types. Occasionally, we receive projects that our partner service labs can't support using standard procedures, necessitating bespoke and innovative solutions.

One example is a recent project in which 392 Fast Technology for Analysis of nucleic acids (FTA) cards were used to collect canine saliva samples and the customer wanted to perform low-pass whole genome sequencing (lpWGS) on the DNA.

Although PCR approaches and amplicon sequencing have been performed, lpWGS of saliva on FTA cards has never been done before. The extremely low DNA yields made this a challenging sample type to work with.

Our service lab tried to extract DNA from the cards, but could not obtain enough DNA to perform our Gencove-modified, miniaturized KAPA HyperPlus library preparation (Roche). To ensure the projects completion, the FTA cards were sent to the Assay Development lab at the Gencove headquarters, where the team was confident we still had options for obtaining adequate DNA to perform lpWGS.

The lab team tested a number of elements of extraction protocols including:

  • The number of FTA card punches to use

  • The extraction buffer conditions

  • The extraction incubation times

  • The elution buffer used for DNA extraction

Punching the cards was not a trivial task. After some experimentation, we found that using 8 punches gave us better results than 4 punches. Punching 392 cards with 8 punches each required 3136 punches, which took a significant amount of time. Furthermore, FTA cards with indicators are normally purple and turn white where the sample is applied, however some cards had very little saliva applied which made it difficult to determine where to punch the cards (Figure 1). Additionally, some cards we received did not have any indicator on them at all, and it was hard to determine the region of the highest concentration of DNA.

Figure 1: Purple indicating and white non-indicating FTA cards with varying amounts of saliva and holes punched to retrieve DNA. Left to right: Good sample with white area indicating the location of saliva, poor sample with little saliva, non-indicating card with saliva difficult to see.

The elution protocol as recommended by Whatman requires incubation for 30 minutes in Tris-HCl at 95C, pulse vortexing 60 times, and elution from the cards. Unfortunately, this approach did not give high enough DNA yields, even with 8 punches, so we tried extracting the DNA with our in-house optimized dried blood spot (DBS) DNA extraction protocol. To achieve the best yields, we attempted incubating the FTA punches in lysis buffer for either 3 hours at 60C or overnight at 60C. Overnight incubation resulted in the degradation of the DNA, while 3 hours of incubation in lysis buffer produced better quality DNA, that was closer to a genomic size profile.

Figure 2: Fragment Analyzer traces of DNA extracted from saliva on FTA cards using the DBS protocol. 3 hours resulted in less damage than overnight incubation at 60C.

The 3-hour incubation DBS protocol produced some samples with high enough yields to undergo the miniaturized KAPA library preparation, however, the majority of samples did not have sufficient material. In fact, some were not readable by our Qubit High Sensitivity reagents (Qubit). So we turned to a low-input kit we have used for other applications; our version of the miniaturized Nextera XT (Illumina).

Our miniaturized Nextera XT protocol only requires 126 pg of starting material. It’s a fast, transposase-based kit and is great for dealing with low sample quantities. We tested it on a subset of the DNA extracted from the FTA cards and found that when eluted from the extraction in Tris-HCl, the resulting libraries were smaller than our desired range. One option to increase the size of the libraries would be to reduce the tagmentation time during library preparation, however since the current time is only 5 minutes, this is logistically difficult so we tried a different approach: eluting in a buffer with some EDTA to attenuate the reaction. We found eluting from the extraction in low EDTA Resuspension Buffer (RSB) resulted in Nextera XT libraries that were slightly larger and a better size for lpWGS.

Figure 3: Fragment Analyzer traces of Nextera XT libraries from canine saliva DNA extracted from FTA cards. DNA eluted without EDTA from the extractions produced very small libraries, while eluting with a low amount of DNA produced good sized libraries.

Once we determined the best extraction method, we prepped all 392 using the 3 hour DBS protocol, eluting in RSB and we got a sufficient amount of library from 100% of samples for lpWGS sequencing. Even the ones without values on the Qubit were able to form libraries and sequence all of the samples.

The data was run through our canine low-pass imputation pipeline and analyzed by our data scientists. Although we observed higher duplication rates than usual (with the first and third quartiles being 22% and 31% respectively), owing to the low input nature of the samples, all but one of the sequenced libraries passed all our internal data QC standards (99.74%) and we were able to generate genome-wide variant calls on all passing samples.

In tackling the unique challenge of sequencing canine saliva samples on FTA cards, we had to get creative. Our goal is always to accelerate our customers' journey from samples to solutions, no matter the sample type, and this project underscores our approach: to explore new avenues and adapt our methods to meet the needs of the projects we undertake.