Which company's NGS library preparation kit for low-input FFPE samples is more reliable?

For low-input FFPE RNA library preparation, I focus on two metrics: the lower input limit and whether SPRI magnetic beads are included. Imported NEBNext Ultra II RNA and the QIAseq series are relatively robust for low-input samples, but lead times are commonly affected by logistics, typically 2–4 weeks; among domestic options, low-input RNA library prep kits from Vazyme, Yeasen, and ABclonal are also used, usually include magnetic beads, and have lead times of about 1 week. I recommend asking distributors for an all-inclusive quote covering "rRNA depletion, adapters, and magnetic beads," rather than comparing only the price of the core reagents per reaction. We generally run 3 replicates each for domestic and imported kits, comparing library yield, fragment size distribution, and Q30 on the same batch of FFPE RNA, which is more reliable than judging by marketing claims alone.

Domestic low-input FFPE RNA library preparation kits now have decent parameters, but you need to keep an eye on batch-to-batch consistency. We've tried two domestic brands. One gave normal library yield at 10 ng input, but after switching batches, the adapter ligation efficiency fluctuated. My advice is not to lock in a single brand all at once. First buy a small pack or request a trial sample, and use the same batch of FFPE RNA to run positive and negative controls simultaneously, checking yield, fragment size distribution, and Q30. When purchasing, ask clearly whether purification magnetic beads, library adapters, and rRNA depletion are included, as many quotes cover only the core reagents. On lead time, domestic suppliers usually have stock available in 3-7 days, while imported kits may take 2-4 weeks if customs clearance is involved. When negotiating bulk pricing, require tax and shipping included and split shipments.

When we work on FFPE transcriptomics, the first step is not to rush into comparing kits, but to take 2-3 sections, extract RNA, and then measure DV200 first. Starting with 10-50 ng total RNA, if DV200 is very low, many low-input kits will also show unstable yield, and the problem may not all be in the kit itself. It is advisable to assess RNA integrity first, and then choose an rRNA depletion strategy or random primer approach according to the degree of degradation. In one of my previous batches, DV200 was clearly low, and only after switching to a library preparation scheme suitable for degraded samples did the library yield become stable. Doing this step first may help you save a lot of trial-and-error costs.

For 10-50 ng FFPE RNA library preparation, polyA enrichment is not recommended. Degraded samples suffer severe polyA tail loss, and the transcriptome will be biased toward the 3' end. Prioritize total RNA library preparation with probe-based or RNase H-based rRNA depletion kits. PCR cycle number also needs key optimization. Low input does not mean the higher the cycle number, the better. FFPE samples have extensive damage, and excessively high cycle numbers tend to cause bias and increased duplication rates. I now typically perform a pre-experiment with amplification curves first to confirm the cycle number at which the plateau is reached, and library yield becomes much more stable. When performing magnetic bead purification, bead drying time and washing must be very consistent; otherwise, recovery rate fluctuates greatly at low input.

Besides the kit itself, operational details have a major impact on the stability of 10-50 ng FFPE samples. After RNA extraction, do not subject it to repeated freeze-thaw cycles; aliquot the RNA into single-use portions before library preparation. Keep the elution volume as small as possible, for example 10-15 μL, to increase the starting concentration. Over-drying purification magnetic beads or residual wash buffer can both affect recovery, and this is especially pronounced at low input amounts. We later assigned one person to perform library preparation consistently to reduce operational variability, and the fluctuation in library yield decreased markedly. Therefore, if a certain imported kit is unstable, first rule out human and sample factors, and then switching kits will lead to clearer conclusions.