* Check RNA-seq read quality with FastQC.
    Version: 0.10.1

* Map reads to genome with Tophat.
  * Build genome index for Tophat using ToxoDB GT1 genome version 9.0.
      Version: bowtie2-intel/2.1.0
      Command: bowtie2-build ToxoDB-9.0_TgondiiGT1_Genome.fasta <bowtie_ref>
  * Map reads with Tophat.
      Versions:
        tophat-gcc/2.0.8
        bowtie2-intel/2.1.0
        samtools-intel/0.1.19
        boost-intel/1.53.0
      Command: tophat -p 8 --output-dir <output_directory> <bowtie_ref> <1.fastq> <2.fastq>

* Check quality of mapping with flagstats
    Version: samtools-intel/0.1.19
    Command: samtools flagstat <accepted_hits.bam>

* Check quality of mapping with RNA-SeQC.
    Version: java/1.6.0_25
    Command: java -Xmx42g -jar /path/to/RNA-SeQC_v1.1.7.jar -r <cleaned_sorted_genome.fasta> -s <list_of_bam_files> -o <output_report> -t <gffread_output.gtf>

* Create gene models with Cufflinks.
    Versions:
      cufflinks-gcc/2.1.1
      tophat-gcc/2.0.8
      bowtie2-intel/2.1.0
      samtools-intel/0.1.19
    Command: cufflinks -p 8 -o <output_directory> <accepted_hits.bam> --overlap-radius 1 --min-isoform-fraction 0.04 --pre-mrna-fraction 0.1

* Merge gene models with Cuffmerge
    Version: cufflinks-gcc/2.1.1
    Command: cuffmerge -p 8 -o <merged.gtf> -s <genome.fasta> <list_of_cufflinks_outs>

* Truncate transcripts to CDS boundaries with GeneGuillotine.
  * Manually inspect reference genome for overlapping genes and remove overlapping fragments. If you run GeneGuillotine first, it will inform you where these are.
  * Run GeneGuillotine.
      Versions:
        GeneGuillotine/1.0
        ruby-gcc/1.9.2
      Command: geneguillotine.rb -i <merged.gtf> -g <genome_no_overlaps.gff> -o <split_merged.gff>

* Analyse differential expression with limma/voom
    Versions:
      limma/3.18.1
      edgeR/3.4.0
      Rsubread/1.12.1
      R/3.0
    Commands:
      t4h_fcounts <- featureCounts(files=t4h_targets$BamFile,file.type="BAM",annot.ext=gff_file,isGTFAnnotationFile=TRUE,nthreads=8,isPairedEnd=TRUE,PEReadsReordering=TRUE)
      t4h_isexpr <- rowSums(cpm(t4h_fcounts$counts) > 3) >= 3
      t4h_x <- t4h_fcounts$counts[t4h_isexpr,]
      t4h_y <- voom(t4h_x,t4h_design,plot=TRUE)
      plotMDS(t4h_y,xlim=c(-2.5,2.5))
      t4h_fit <- eBayes(lmFit(t4h_y,t4h_design))
      topTable(t4h_fit,coef=2)
      write.csv(topTable(t4h_fit,coef=2,number=Inf,p.value=0.05),file="/tmp/limma_out_4h")

* Analyse differences in alternative splicing with DEXSeq.
    Versions:
      DEXSeq/1.8.0
      python-gcc/2.7.5
      R/3.0
    Commands:
      * Modify dexseq_prepare_annotation.py to work in unstranded mode, by changing
          exons = HTSeq.GenomicArrayOfSets( "auto", stranded=True )
            to
          exons = HTSeq.GenomicArrayOfSets( "auto", stranded=False )
      * Create flattened GFF.
          python dexseq_prepare_annotation_unstranded.py <input.gtf> <output.gff>
      * For each bam file,
          samtools view -o 8_2_v2.sam 8_2_v2.bam
          sort -k1,1 -k2,2n 8_2_v2.sam > 8_2_v2_sorted.sam
          python dexseq_count.py --p yes -r name -s no <flattened.gff> <inputsam> <outputfile>
      * Import into R, then
          ecs0_vs_4 = read.HTSeqCounts(countfiles = file.path(inDir, paste(rownames(samples))), design = samples, flattenedfile = annotationfile)
          sampleNames(ecs0_vs_4) = rownames(samples)
          ecs0_vs_4 <- estimateSizeFactors(ecs0_vs_4)
          sizeFactors(ecs0_vs_4)
          ecs0_vs_4 <- estimateDispersions(ecs0_vs_4, nCores=8)
          ecs0_vs_4 <- fitDispersionFunction(ecs0_vs_4)
          ecs0_vs_4 <- testForDEU( ecs0_vs_4, nCores=8)
          ecs0_vs_4 <- estimatelog2FoldChanges( ecs0_vs_4 )
          res_0_vs_4 <- DEUresultTable(ecs0_vs_4)
          table ( res_0_vs_4$padjust < 0.05 )
          table ( tapply( res_0_vs_4$padjust < 0.05, geneIDs(ecs0_vs_4), any ) )
          plotMA( ecs0_vs_4, FDR=0.5, ylim=c(-4,4), cex=0.8 )
          DEXSeqHTML( ecs0_vs_4, FDR=0.05, color=c("#FF000080", "#0000FF80") )

* Analyse alternative splicing with JunctionJuror.
    Versions:
      JunctionJuror/1.0
      ruby-gcc/1.9.2
    Commands: junctionjuror.rb -j <junctions.bed.list> -g <genome_no_overlaps.gff> -o <output_genes> -t 3