Blog @ Illumina
Real scientists. Real commentary.

Streamlining Next-Generation Sequencing Experiments with NeoPrep Digital Microfluidics

Lisa C. Watson and Gary P. Schroth, Illumina Applications Group
| Feb 27, 2015

This week, Illumina launched the NeoPrep System – the first Illumina digital microfluidics platform for automated preparation of sequencing libraries. Digital microfluidics uses electrical voltage to manipulate nanoliter volume droplets through standard library prep chemistry, to ultimately transform sheared DNA or total RNA into ready-to-sequence libraries. This revolutionary technology enables 16 libraries to be prepared in parallel, with only 30 minutes of initial hands-on time, freeing up your schedule for other projects or planning your next experiment. The NeoPrep platform is perfect for researchers who are ready to increase sample throughput, simplify their workflow, or accelerate their RNA-Seq, transcriptome profiling, or whole genome sequencing experiments.

If you are familiar with NGS library prep, you know that you need to plan your day around critical wash and incubation steps. Now, with the tap of a touchscreen, samples are concentrated into nanoliter volumes and merged with droplets containing the components for executing each sequential enzymatic reaction with exquisite precision. 

One of the key features of the NeoPrep system is the capability to quantify and normalize final libraries, offering a complete solution that eliminates the need for additional assays, instruments and calculations before loading your libraries on a sequencer. By standardizing library quantification with the NeoPrep platform, it’s even easier to hit the optimum target cluster density, since we’ve already determined the seeding concentration for each of our sequencing platforms.

Figure 1:  (Click to enlarge) Uniform CV of index representation of normalized NeoPrep mRNA libraries.  Index distribution of sixteen normalized TruSeq Stranded mRNA libraries sequenced on a HiSeq 2500 flow cell.

In developing the NeoPrep system, one of our goals was to be able to use lower sample input than our manual kits. The manual TruSeq Stranded mRNA kit supports a minimum of 100 ng total RNA input. With the efficiency of digital microfluidics, the NeoPrep system delivers equivalent or superior quality libraries from as little as 25 ng input.  During development, we tested the limits of the NeoPrep system and were pleased to find that we could generate high quality libraries that meet most key performance metrics and correlate well with manual preps from less than 25 ng total RNA input. Even at 2 ng, the quantification of gene expression counts corresponds with manually prepared libraries that started with 50-fold more input of total RNA (Figure 2).

Figure 2: (Click to enlarge) The NeoPrep system delivers high-quality mRNA libraries from low input amounts.High correlation of gene-level expression in fragments per kilobase of transcript per million fragments mapped (FPKM) [Trapnell et al. (2010), Nature Biotechnology, 28, 511-515] is demonstrated for NeoPrep and manually prepared mRNA-seq libraries. As expected, the number of sequencing duplicates increases at very low RNA input amounts, as the number of unique starting molecules becomes limiting. 

As part of Illumina’s Applications group, we recognize that converting biological samples into sequencing libraries is only the first step, and that the ultimate goals are discoveries stemming from insightful interpretation of NGS results. Over the last few years, Illumina has focused on introducing products that both shorten and simplify the sample-to-answer process. Last year we introduced the NextSeq 500 sequencer, which offers easy single-cartridge loading and overnight run times and generates over 400 million paired-end reads per run. As for data analysis, if you’re not a bioinformatics expert, we’ve got that covered too. BaseSpace cloud computing solution enables sequencing data to stream directly from any of Illumina’s sequencing platforms to our cloud analysis environment, where you can launch a variety of easy-to-use analysis apps and be looking at your data within hours.

Table 1: Steps and Hands-On Time for NeoPrep

We put our streamlined sample-to-answer workflow to the test by combining NeoPrep, NextSeq and BaseSpace in a standard RNA-Seq application: gene expression profiling and fusion detection in cancer samples. Total RNA from 8 commonly-studied cancer cell lines (in replicate for 16 total), was loaded on the NeoPrep at the end of the workday on Monday. Since the complete process takes less than 11 hours, it was done the next morning. In fact, even before leaving for work on Tuesday, the library prep results were waiting in my inbox with a link to a report in BaseSpace. All samples had passed minimal yield expectations and were already normalized to 10 nM concentrations, ready for loading on an Illumina sequencer.

Figure 3: (Click to enlarge) Parallel library preparation on NeoPrep of RNA derived from cancer cell lines. The BaseSpace report generated from preparing 16 NeoPrep TruSeq Stranded mRNA libraries from 8 cancer cell lines (in replicate) is shown.

The NextSeq 500 sequencer has ideal throughput for gene expression profiling experiments downstream of NeoPrep. Sixteen RNA-Seq libraries will yield approximately 25 million paired-end reads per NextSeq sequencing run, a sufficient read depth for many RNA-Seq applications. The sixteen libraries were pooled and loaded onto a NextSeq sequencer during the day on Tuesday, taking less than an hour.  We ran a 2 X 75 bp paired-end run, as we have found this to be the optimal read length for TruSeq RNA libraries, and set the NextSeq system to stream the data directly to BaseSpace. Since this entire run takes less than 16 hours to complete, the data were visible early Wednesday morning. 

Before leaving the house on Wednesday we were able to spend 5 minutes to start the TopHat Alignment App in BaseSpace to begin processing this data. The TopHat App is used to align reads, quantify gene expression, and call variants. Additionally, the TopHat App enables detection of novel gene fusions, providing a straightforward pipeline for examining cancer translocations using RNA-Seq.  By late afternoon on Wednesday, we were able to access a list of candidate fusions for each sample along with their precise breakpoints. For example, 39 reads supporting an EGFR fusion are detected in the A431 cell line (Table 2). In about 48 hours (late afternoon on Monday to late afternoon on Wednesday) we had gone from total RNA to fusion transcript discovery – with only about 90 minutes of total effort!

Table 2 (Click to enlarge): Sixteen normalized NeoPrep TruSeq Stranded mRNA libraries were pooled, denatured and sequenced on a NextSeq 500 at 2 x75 bp paired-end reads. Fusion analysis was performed using the BaseSpace TopHat Alignment App. The number of fusion supporting reads detected (combined between replicates) for selected gene fusions from several cancer cell lines is shown.

If your study aims to compare expression differences between two distinct sample sets, the Cufflinks Assembly and DE App makes it easy to obtain a list of statistically significant genes.  We compared gene expression profiles of acute promyelocytic leukemia cells (HL-60) and acute monocytic leukemia cells (THP-1).  An interactive plot can be used to view the gene list and filter results by the magnitude of fold-change differences in gene expression.  By viewing read alignments in a genome browser, you can visually confirm that transcript coverage across your sample types is consistent with Cufflinks results (Figure 5). 

Figure 4: (Click to enlarge) Combining NeoPrep, NextSeq, and BaseSpace to detect differentially expressed genes. Scatterplot generated by the BaseSpace Cufflinks Assemby and DE App comparing gene expression levels, in log2(FPKM), for RNA from HL-60 and THP-1 cell lines. The Integrative Genome Viewer screenshot shows transcript coverage of the MYC gene, indicating higher expression of MYC in HL-60 cells compared to THP-1 cells. Tracks are scaled equivalently (normalized by read depth).

In conclusion, the NeoPrep system offers a simple walk-away solution for library preparation that has inspired our Applications group to re-envision how we do NGS experiments.  The low input capabilities of the NeoPrep system enable precious samples that were previously not ideal for NGS to become the focal point for new studies.  The time saving gained by automating the library prep process lets you focus on what really counts: interpreting the data.