2.6 | Differential gene expression analysis
Differential gene expression
(DGE) analysis was conducted using the program DESeq2 version 1.22.2
(Love et al., 2014). The DESeq2 package has a normalization function
implemented based on the median of ratios method, in which the geometric
mean of the gene counts across all samples is used to calculate the
ratios of each gene and each sample, allowing between-sample comparisons
(hbctraining, DGE_workshop (2022), GitHub repository, [accessed
05.04.2022]; https://github.com/hbctraining/DGE_workshop).
Additionally, a variance stabilizing transformation (VST) was performed
on the data to remove variance-mean dependence (Anders & Huber, 2010).
All genes were taxonomically assigned with MEGAN6 version 6.13.1 (Huson
et al., 2007); only ascomycete, chlorophyte and cyanobacterial genes
were retained for DGE analysis (after DGE analysis the term ’genes’ will
be used instead of ’transcripts’ to be congruent with the terminology of
’differentially expressed genes’). The vst-normalized data of each of
the three taxonomic units was used to perform Principal Component
Analysis (PCA) (R version 3.5.2). In the DGE analyses, we quantified
differences in fungal gene expression owing to morph type (tripartite
vs. cyanobacterial) and those in fungal, algal and cyanobacterial gene
expression owing to temperature. Transcripts with an adjusted
Benjamini-Hochberg p -value < 0.05 and a
log2-fold change > |2|
were regarded as significantly differentially expressed. Our analyses
focus on the 200 most significantly differentially expressed genes as
determined with two-way ANOVAs for all organisms. Functional annotation
of these differentially expressed genes was conducted using UniProt
BLAST (The UniProt Consortium, 2021). The BLAST search was run using
default settings with the target databases being “Fungi”, “Plants”
and “Bacteria”. The best alignment based on e-value
(<10-5) was used to infer gene functions.
The top-200 fungal differentially expressed transcripts were also
blasted (blastx version 2.7.1+, translated nucleotide to protein)
(Sayers et al., 2020) against our own database consisting of filtered
metagenomic sequences of Peltigera britannica , P.
leucophlebia and P. collina (unpublished data of the authors)
using standalone BLAST for Linux Ubuntu (ncbi-blast+ package). This
latter step was carried out to evaluate if the differentially expressed
ascomycete genes were likely to originate from the lichen mycobiont or
from other lichen-associated fungi. In the former case, there should be
a hit both in the P. britannica metagenome, and in at least some
of its congeners. The P. britannica metagenome was sequenced from
a lichen individual not included in transcriptome sequencing
(unpublished data by Werth, Andrésson, Resl and Warshan) and was built
after de novo transcriptome assembly and DGE analysis. Gene
Ontology (GO) annotations of all DEGs were conducted with the
Bioconductor package topGO version 2.34.0 (Alexa & Rahnenführer,
2018).