other plots are realized by the “RColorBrewer” R package
(https://cran.r-project.org/web/packages/RColorBrewer/
RColorBrewer.pdf).3 Methods
In this section, we will show how to analyze RNA-seq expression
data using iSeq. We reanalyzed a public RNA-seq data from Gene
Expression Omnibus (GEO accession: GSE39866), which was
published onNature Neuroscience[27]. This dataset reported the
mRNA expression differences between embryonic and adult mouse
cerebral cortex and identified several genes (Mobp,Igf2bpet al.)
involved in important pathways (ion transport, cell cycle) with
altered expression during the cerebral cortex developmental pro-
cesses. Using iSeq, we could easily reproduce the same results as in
the paper without using command line analysis or programming.3.1 Prepare Data In theNature Neurosciencepaper, the authors generated seven
RNA-seq samples for two biological conditions, consisting of four
replicate samples for embryonic and three replicate samples for
adult mouse cerebral cortex.
- Expression File of Genes
Download the expression data file fromhttp://202.205.131.
33:3838/expression.csv. In this file, each row represents a
certain gene, and each column represents a certain sample
(Fig.2). The value of each entity represents the expression
level as measured by raw sequence read counts; the FPKM
and TPM are also acceptable. The first column and row list
the names of genes and samples, respectively. - Condition File of Samples
Download the sample description file fromhttp://202.205.
131.33:3838/condition.csv. This file allows classifying samples
into biological conditions. It has two rows, with the first row
listing sample names and the second row listing condition
names (Fig.2). Make sure to use the same sample names in
the expression file and condition file.
3.2 Upload Data There are two ways to access iSeq: either visit the online version
(http://iseq.cbi.pku.edu.cn) or install a local version of iSeq on
your own computer (https://github.com/ChengLiLab/iSeq).
Figure2 shows the interface to upload the expression table (expres-
sion.csv) and the sample condition table (condition.csv).
3.3 Normalization Some systematic variations have been reported, such as the library
size, gene length, and GC content [28]. Normalization is an import
step for RNA-seq to remove these systematic biases. After
172 Chao Zhang et al.