RNA-Seq Experiments and Data Filtering

Leaves were collected from field-grown plants at the end of the experiment (August 4, 2016). The leaf samples were immediately soaked in an RNA preservation buffer of pH 5.2 consisting of 5.3-M (NH₄)₂SO₄, 20-mM EDTA, and 25-mM trisodium citrate dihydrate at 4°C overnight and stored at −80°C until RNA extraction. Total RNA was extracted using the Maxwell 16 Lev Plant RNA Kit (Promega Japan, Tokyo) according to the manufacturer’s protocol. Selective depletion of rRNAs and highly abundant transcripts was conducted prior to RNA-Seq library preparation as previously described (Nagano et al., 2015). Then, RNA-Seq libraries were prepared as previously described (Ishikawa et al., 2017). Sequencing using Illumina HiSeq^® 2500 was carried out by Macrogen Co. We sequenced 92 samples per lane and obtained 829,681 mapped reads per sample on average.

The fastq files generated by sequencing were preprocessed using Trimmomatic version 0.32 (Bolger et al., 2014). The preprocessed sequences were mapped onto the A. thaliana reference genome (TAIR10 cDNA) using Bowtie version 1.1.1 (Langmead et al., 2009) and then quantified using RSEM version 1.2.21 (Li and Dewey, 2011). The parameter settings of Trimmomatic, Bowtie, and RSEM were the same as those described by Kamitani et al. (2016). Following Kamitani et al. (2016), we calculated the raw read counts and reads per million (rpm) from the expected read counts generated with RSEM. Transposable elements were excluded prior to statistical analyses. We calculated the total raw read counts for each plant sample and discarded shallow-read samples belonging to the lower fifth percentile of the total raw read counts ( Figure 1C ). Consequently, samples with >12,130 reads were subjected to statistical analyses. To exclude non-expressed genes, we then averaged log₂(rpm + 1) for each gene between all plant samples and eliminated genes with an average log₂(rpm + 1) of zero ( Figure 1C ). Overall, we obtained a final dataset on 24,539 genes for 173 plants. In this final dataset, 53 out of 173 samples had <10⁵ total reads. Overall trends did not change when we set the threshold at 10⁵, although the statistical power decreased due to lower sample size.

Sequence data from our RNA-Seq were submitted to the NCBI Sequence Read Archive repository under the BioProject number, PRJNA488315 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA488315). Read count data and source code are available via the GitHub repository (https://github.com/naganolab/AthRNAseq2016Zurich_Sato_et_al).