RNA-seq Analysis Portal – Simplify Your RNA-seq Insights

From Illumina, the TruSeq Stranded Total RNA Library Prep (Human/Rat, Gold, Globin) AND the Stranded Total RNA Prep with Ribo-Zero Plus.

From New England Biolabs, the NEBNext UltraTM II Directional RNA Library Prep Kit for Illumina Roche Sequencing solutions.

From KAPA / Roche, the KAPA RNA HyperPrep Kit.

From Takara Bio, the  SMARTer Stranded Total RNA Sample Prep Kit - HI Mammalian AND the SMARTer Stranded Total RNA Sample Prep Kit - Low Input Mammalian.

From Thermo Fisher Scientific, the Collibri Stranded RNA Library Prep Kit for Illumina Systems.

Human (Homo sapiens), Mouse (Mus musculus), Rat (Rattus norvegicus)

Cow (Bos taurus), Dog (Canis lupus familiaris), Horse (Equus caballus), Crab-Eating Macaque (Macaca fascicularis), Rhesus Macaque (Macaca mulatta), Rabbit (Oryctolagus cuniculus), Pig (Sus scrofa)

Fruit Fly (Drosophila melanogaster), Roundworm (Caenorhabditis elegans), Zebrafish (Danio rerio), Thale Cress (Arabidopsis thaliana), Baker's / Brewer's / Budding Yeast (Saccharomyces cerevisiae), Fission Yeast (Schizosaccharomyces pombe)

Chicken (Gallus gallus), Rainbow Trout (Oncorhynchus mykiss)

No, the RNA-seq Data Analysis Portal cannot analyze single-cell data. However, it is designed to support relative gene expression profiling from larger samples such as bulk cells in culture and tissues.

RNA sequencing data is analyzed in three steps: primary, secondary and tertiary data analysis. Primary RNA sequencing analysis converts sequencing reads into gene-specific raw expression values. Secondary RNA sequencing analysis calculates fold-change and p-value results from raw expression values. Tertiary RNA sequencing analysis interprets those results through comparisons with known pathways and diseases and by identifying upstream regulators.

The general steps performed by primary RNA sequencing analysis pipelines include, but are not limited to:

1. Demultiplexing sample indices if used to multiplex samples in the same sequencing run
2. Correcting for amplification errors by merging sequencing reads into Unique Molecular Index (UMI) reads
3. Trimming reads of low quality and ambiguous nucleotides as well as any adapter sequences if needed
4. Mapping sequences to the correct species-specific transcriptome to identify the sequenced genes
5. Counting the number of UMIs per gene to obtain their raw expression values in each sample

However, with the RNA-seq Analysis Portal and its fixed but optimized RNA sequencing analysis pipeline, you don’t have to worry about these details. The application does it all for you. As such, biological researchers like yourself can focus on the critical parameters of secondary differential gene expression analysis.

After completing primary RNA sequencing analysis to obtain raw gene expression values for each sample, secondary differential gene expression analysis then:

1. Normalizes the raw data to control for variability in performing the experiment from RNA isolation through the sequencing runs
2. Averages the normalized data across the same groups of replicate samples
3. Calculates the fold change values or the ratio of the average normalized gene expression values between two experimental groups
4. Calculates p-values to determine the statistical significance for each fold change value

As easy as these calculations are to perform, the RNA-seq Analysis Portal automatically performs all them for you, once you set up your RNA sequencing-based differential gene expression analysis by defining the samples and groups to be compared. From there, the fold-change and p-values and other intermediate values are used for RNA-seq data visualizations such as a volcano plot and a heatmap, again automatically generated by the RNA-seq Analysis Portal for you.

Once determined, a list of statistically significant differentially expressed genes is interpreted by comparing with known pathway and disease gene lists and by identifying upstream regulators.

The list of statistically significant differentially expressed genes are compared with genes known to be involved with specific pathways or associated with specific diseases. Such a comparison helps biological researchers like yourself gain a deeper understanding of biology behind, or driven by, the differential gene expression observed in the experiment. Pathway comparisons can also even help you gain a deeper understanding of how the differentially expressed gene are regulated.

The number of differentially expressed genes, the number of genes associated with the pathway or disease, and the number of overlapping genes all define the statistical significance and the relevance of that overlap. The result of such enrichments are then typically sorted by decreasing statistical significance to bring the most relevant ones to the top.

Collecting all the genes for all the possibly relevant pathways and diseases can be extremely time consuming, and not all of this information is publicly available. The statistical significance calculations can also be rather complicated.

However, the RNA-seq Analysis Portal automatically compares your list of differentially expressed genes with comprehensive gene lists in a knowledge base containing a comprehensive list of pathways and diseases. The RNA-seq Analysis Portal also automatically calculates the statistical significance of those overlaps and returns a list of the top ten most relevant pathways and the top ten most relevant diseases, Finally, it even reports the top ten most relevant predicted upstream regulators including miRNA.

Normalization should be performed in a way that best controls for variability in the experimental protocol from the RNA isolation step through the sequencing runs.

A normalization factor or value is calculated for each sample that should be roughly uniform across the analyzed samples’ datasets. The raw expression values are then divided by the sample-specific normalization factor before continuing with averaging the normalized data and calculating fold change values and p-values.

Various normalization methods exist, and normally you would try a few and choose one that seems to work the best for your dataset. However, with the RNA-seq Analysis Portal and its published and well-established normalization method, you don’t have to worry about these details. It already uses a published and widely-accepted normalization method.

Any good experimental setup must include replicate treated or affected experimental samples and replicate untreated or unaffected control samples for comparison.

A critical factor is setting up any differential gene expression analysis is to define the groups for comparison. Any good experimental setup includes not only a set of treated or affected experimental samples or test samples, but also a set of untreated or unaffected control samples to which all experimental samples or test samples are compared.

Datasets obtained from samples using the same biological condition are therefore grouped and their normalized data across each individual gene averaged before calculating fold-change values during the secondary differential gene expression analysis. Experimental and/or biological replicates for every condition are also crucial for determining the statistical significance of those differential gene expression analysis results by calculating p-values.

Once these group definitions have been made in the RNA-seq Analysis Portal, the application takes care of the rest of the differential gene expression analysis for you.

Fold change vales are the ratios of the average normalized expression values between experimental and control groups for each gene in the dataset.

Fold change values greater than one indicate up-regulation in experimental groups relative to control groups. Fold change values less than one indicate down-regulation in experimental groups relative to control groups.

As easy as these calculations are to perform, the RNA-seq Analysis Portal automatically calculates them for you, once you set up your RNA sequencing based differential gene expression analysis by defining the samples and groups to be compared.

The p-values represent the likelihood of obtaining the observed results if no real difference existed, and various methods are available to calculate p-values.

The lower the p-value, the greater the statistical significance of the associated fold change values. The higher the p-value, the smaller the statistical significance of the associated fold change values. More extreme fold change values tend to have smaller p-values. More samples can improve p-values, especially for smaller changes in expression or smaller fold change values.

Calculating p-values is a bit complicated, and there are a number of methods to correct them for false discoveries. However, the RNA-seq Analysis Portal automatically calculates them for you and corrects for false discoveries, once you set up your RNA sequencing-based differential gene expression analysis by defining the samples and groups to be compared.