Transcriptomics: RNA sequence and microarray data searches

Learning Objectives

● Review the types of expression searches in VEuPathDB

● Use the differential expression, fold change and percentile searches to explore gene expression
in liver stage plasmodium infections
● Compare the expression searches to reveal advantages and disadvantages of each search
● Run a co-expression search.

Transcript expression or the abundance of an mRNA, can be determined in the laboratory with several
different techniques including RNA-sequence, microarray, and RT-PCR. VEuPathDB supports these data
types with several searches (see table below). For RNA sequence data, expression values are graphed
on gene pages and mapped reads can be visualized in the genome browser. Using the search strategy
system, it’s easy to delve deep into a specific data set and to take advantage of several types of data
when combining search results in the strategy system.

RNA- Micro-
Search Description
seq array
Differential Statistical analysis of studies whose experimental design includes
Expression biological replicates. A differential expression search finds genes
based on fold change difference between two samples with a user ✔
defined p-value cutoff. Only pairwise comparisons can be made
with this search.
Fold Change Expression differences between samples are calculated but
statistical analyses are not performed. A fold change search finds
genes whose expression value differs between samples without
considering statistical parameters. This search offers a form of
✔ ✔
differential expression analysis when the experimental design did
not include replicates and allows for comparing groups of samples,
e.g. find genes whose expression is up-regulated in the liver time
course (2, 24, 36, and 54 hours) vs the control (0 hours).
Percentile For each sample in an experiment, each genes’ expression value is
sorted from lowest to highest and a percentile rank is determined.
✔ ✔
For example, a percentile search can find genes whose expression
is in the highest 10% of expression values within a sample.
Sense/Antisense For strand-specific RNA sequence, expression values are
determined in the sense and antisense direction. This search finds
genes that exhibit simultaneous changes in sense and antisense
✔
transcripts. For example you can look for genes with increasing
antisense transcripts and decreasing sense transcripts, as might
occur when antisense transcription suppresses sense transcription.
Splice-site This trypanosome-specific search takes advantage of the ‘splice-
Location leader’ RNA-seq data which determines transcript abundance ✔
within the polycistronic mRNA using splice-leader specific primers.
 This search identified genes whose 5' splice site location varies
between samples.
Metacycle The MetaCycle package detects rhythmic signals from large scale
time-series data, such as circadian rhythms within expression time
courses, using either ARSER or JTK-Cycle. This search returns
genes whose rhythmic signals match the conditions (period and ✔ ✔
amplitude range) you specify. The search will return the
corresponding period, amplitude and p-value of genes that meet
your search criteria.
Similarity The similarity search returns genes whose expression profile
within the experiment follow a similar pattern as the gene you ✔ ✔
specify.
Direct Microarray data for two samples is often collected on the same
Comparison glass slide. For these experiments, the direct comparison search
✔
returns genes whose expression varies between samples in
pairwise comparisons.
Coexpression Meta-analysis across multiple microarray experiments defined a
co-expression network. This search returns genes within the co- ✔
expression network of your gene(s) of interest.

1. Find genes that are up-regulated in the later liver stages of Plasmodium infection. [Link]

The life cycle of Plasmodium is split between the sexual

mosquito stage and the asexual host phase. The host
stage includes a 6-7-day asymptomatic liver stage which
ends with the release of merozoites into the bloodstream
where they infect erythrocytes. The erythrocytic stages
are well studied compared to the liver stages.

PlasmoDB contains RNA seq data from a study in the

rodent model Plasmodium berghei, that includes a time
course of liver infection as well as sporozoite and
merozoite samples for comparison. (Caledlari et al. 2019)
Seven samples were assayed in triplicate for RNA
sequence:
1. Sporozoites
2. 6 hr liver infection
3. 24 hr liver infection
4. 48 hr liver infection
5. 54 hr liver infection
6. 60 hr liver infection
7. Merozoites (detached cells).

Use this data set to determine what genes are upregulated at least 4 fold (p-value <= 0.001) at 48
hr post infection vs the sporozoite stage.
a. Navigate to the RNA seq search page and find the data set called Ex-erythrocytic stage
transcriptomes (sporozoite, liver time course and detached cells) (Caldelari et al.). Searches are
available from the Search For… menu on the left side of the home page, as well as the Searches drop
down menu in the header.

b. Arrange the differential expression search to return genes that are at least 4 fold up-regulated in the
48-hour liver infection compared to sporozoites with a p-value of p<0.001.
c. How many genes were returned by the search? Do you believe these results? To convince
yourself, you could browse the product description column. Are there clues that these genes are
liver-specific?
d. Increase the statistical stringency of the search from p <= 0.001 to p<0.0001. How many genes are
returned by the search now? Hint: revise the search and change the p-value. Hover over the yellow
search box until the Edit icon appears. Click the Edit icon and choose revise from the options panel.

e. What other properties would you expect of a late liver stage gene/protein? Since the next step is to
emerge from the hepatocyte, these genes may have proteolytic activity. Intersect your RNA seq
search with a GO term search to see if any of your genes are annotated with proteolytic or
peptidase activity. (GO:0008233 peptidase activity GO:0006508 proteolysis ) How many genes have
these activities?
2. Find genes that are upregulated 4 fold in any liver stage compared to sporozoites. Hint: use the
Fold change search to compare the 6, 24, 48, 54 and 60-hour time points to sporozoites.
a. Navigate to the RNA Seq search page and choose the Fold Change search for the Ex-erythrocytic
(Caldelari et al 2019) data set as in 1a above.
b. Arrange the fold change search to return genes that are up-regulated in the average expression
across the liver stages compared to the sporozoites.
c. Explore your results. Did the search return more genes or fewer genes than the differential
expression search?
d. Use the Add Columns to turn on the TPM graph for the ‘Ex-erythrocytic stages’ data set. Notice the
error bars for the DNAJ protein PBANKA_1203800. Would this gene be returned by the Differential
Expression search that applies statistics before returning genes?
e. Use the Percentile search to determine what genes in this result are also expressed in the top 10%
of genes in the merozoite (detached cells) sample? Hint: Add a step to the strategy that intersects
your current result with search that returns the 90-100th percentile genes of the merozoite sample.
3. Find Aedes aegypti genes that are upregulated in both head and muscle during infection with
Wolbachia. The Wolbachia strain wMelPop, which reduces longevity in Drosophila melanogaster,
has been introduced into the Dengue virus mosquito vector, Aedes aegypti as a strategy to reduce
disease transmission. VectorBase has a microarray data set that compared Wolbachia infected and
uninfected mosquito head and muscle. This exercise uses [Link].

a. Navigate to the microarray search and choose the Direct Comparison search for the dataset titled
‘Infection with a Virulent Strain of Wolbachia Disrupts Genome Wide-Patterns of Cytosine
Methylation in the Mosquito Aedes aegypti (Ye et al.)’

b. Initiate a search that returns genes that are upregulated 2 fold in infected head vs uninfected.
c. Intersect your search result with another search that returns genes upregulated 2 fold in muscle vs
uninfected. Your combined result will be genes that are upregulated in head and muscle in response
to Wolbachia infection.
d. Determine enriched Molecular Function GO terms for the upregulated genes. Make sure you are
viewing the combined result (the Step 2 result will be highlighted yellow) and click Analyze Result to
open the Enrichment Tool. What gene functions are shared by the combined result? What
biological role can you envision for these mosquito genes during the wolbachia infection?

4. Find genes that are likely co-expressed with An04g07430, an Aspergillus niger protein coding gene
with little functional annotation. By finding genes that are expressed at the same time as
An04g07430, we may find clues about its function and the biological processes that it participates in.
This exercise uses FungiDB.
a. Navigate to the microarray searches in FungiDB and choose the Coexpression search for the data set
titled Aspergillus niger gene co-expression network (Vera Meyer). Schape et al Nucleic Acids
Research 2019. This data are the results of a meta-analysis of 155 publicly available transcriptomics
analyses for A. niger, which were used to generate a genome-level co-expression network and sub-
networks for >9,500 genes.
b. Run the search to find the co-expression network for An04g07430.

c. What genes share the co-expression profile of An04g07430? Several genes have a correlation
coefficient of 0.85. What are these genes? Visit their gene pages to learn more.
d. Scan the product description column for genes with known functions. Use the Column Histogram
tool to view a word cloud of the product descriptions in the column. Set the rank range to 25-50.
What words occur most often in the product descriptions of An04g07430 co-expressed genes?
e. Run the enrichment analyses for Molecular Function, Cellular Component and Biological Processes.
Do these provide information about what this group of co-expressed genes might be doing?