Optimizing RNA-Seq Workflows

RNA sequencing (RNA-seq) is an instrumental tool for the high-throughput study of gene transcription and its regulation. Despite its widespread use, and the emergence of user-friendly prep kits, RNA-seq workflows are rarely smooth sailing. Common challenges include securing sufficient high-quality sample, depletion or enrichment strategies, and finding data analysis tools. This article offers guidance for overcoming such challenges, and practical tips for optimizing RNA-seq workflows.

Sufficient high-quality sample

Sometimes simply obtaining enough high-quality sample to perform RNA-seq presents a challenge. Several options can help. First, starting with sample collection, minimize RNA degradation upfront. This includes creating a workspace free of RNase, and preserving samples using a nucleic acid protectant to inhibit RNase, or flash freezing. Keeping samples (or, later, extracted RNA) as cold as feasible at every step is important for minimizing degradation. This can be assured by storing at -80 ℃, and by keeping it on ice during use. In addition, “minimizing freeze-thaw cycles of RNA is also a good practice, so if samples are chosen for multiple experiments, it is a good idea to aliquot them,” recommends Keerthana Krishnan, Development Scientist at NEB.

Testing sample quality is important to avoid wasting time and resources on low-quality samples—especially from sources prone to degradation, such as formalin-fixed paraffin-embedded (FFPE) samples. “Despite any efforts to prevent further degradation of RNA, some RNA samples may have too low a quality to continue,” says Bellal Moghis, Director of NGS Product Marketing at Agilent Technologies. “It is important to assess the quality of your sample using reliable, automated electrophoresis systems, such as the Agilent TapeStation or Fragment Analyzer systems.”

Protocol changes may also boost sample quality. “If experimental design allows, individual RNA purification preparations can be pooled until a sufficient quantity is collected,” says Lindsey Cambria, Support Team Manager at Roche Sequencing and Life Science. “If this isn't possible, I would recommend exploring the different commercially available RNA library prep kits and testing those that offer low-input workflows.”

Many column-based extraction kits result in high yields of purified RNA. But if your samples are very low input or even single cells, you may find greater success with a specially designed RNA library-prep kit. “Some library-prep kits skip the RNA isolation step, such as our Single Cell/Low Input RNA library-prep kit,” says Krishnan. “A researcher can choose to extract total RNA, or they can choose to sort cells directly into the cell lysis buffer and proceed directly into library prep, giving them both flexibility and the opportunity to save some time.”

Data analysis tools

The widespread use of RNA-seq means that data analysis tools are increasingly easier to find. Many commercially available NGS library-preparation kits come with data analysis software, though open-source software is another option for researchers with a bioinformatics background. “Commercial software providers such as Partek have enabled even novice users to analyze large NGS data sets using a user-friendly interface,” says Moghis. “Partek Flow automatically processes and analyzes RNA-seq data and then passes results seamlessly to the Agilent Alissa Interpret platform, where it is combined into a single user-friendly report.”

Moghis notes that RNA sequencing workflows can have additional considerations compared to DNA sequencing workflows. “For example, it is important to consider how many bases you need to sequence into your library and how much sequencing depth will be required to properly identify the library fragment, align it to the reference transcriptome, and count the number of reads to assess the expression level,” he says. Standard RNA-seq guidelines—for example, from the ENCODE consortium—can help you to determine the amount of sequencing needed.

Krishnan says that “the best way to find the appropriate tools for analysis will be to find peer-reviewed publications that have similar experimental setups or biological questions.” In addition, she recommends Galaxy—an open, web-based platform for computational research—for helpful information on RNA-seq data analysis.

Further solutions

The challenges posed by degraded or low-input samples can be overcome with other strategies, such as “by modifying adapter concentrations, post-ligation clean-up ratios and/or PCR cycles,” says Cambria. For degraded samples, beginning with a greater mass of input material may also help.

Reducing the number of cleanup steps in the library-prep workflow can also improve RNA-seq results. “Library-prep kits that have been optimized to remove or combine cleanups are quicker and less error-prone than kits requiring many rounds of cleanups, which inherently cause some loss of library, so fewer cleanups translates to higher yields,” says Krishnan. “These streamlined kits, like our Ultra II RNA kits, are also easier to automate, because they require fewer steps and less tip usage.”

Other considerations for degraded samples include the depletion of rRNAs, or the enrichment of mRNAs, prior to library preparation (if applicable). For example, if input RNA is degraded, there is likely a better chance of success with an rRNA depletion workflow, when compared to an mRNA-enrichment workflow. “Many mRNA-enrichment strategies rely on an intact polyA-tail,” notes Cambria. “Degraded RNA samples are less likely to have intact transcripts, making this a less-than-ideal choice for enrichment.”

Converting low-input samples from less stable RNA to more stable cDNA as soon as possible—even before ribosomal RNA (rRNA) depletion—can be advantageous. The traditional sample-handling steps involved in rRNA depletion may result in sample loss that can be significant when studying rare RNA. “If you choose a depletion strategy that removes cDNAs corresponding to rRNA after you have created a library, you stand a better chance of retaining small and rare transcripts, and RNA from limited or degraded samples, providing a truly more holistic view of the transcriptome,” says Steven Blakely, Director of Gene Expression Business and Life Science Research Software Strategy at Bio-Rad Laboratories.

Finally, it never hurts to seek advice from more experienced researchers, such as core NGS facilities at academic research institutions. “These core labs often serve as a great resource to better understand how RNA-seq can be leveraged for your specific experiment,” says Moghis. “By leaning on core labs, or commercial service providers for those that don’t have access to these labs in their institutions, new users can be expertly guided.”