Extracting DE transcripts and generating heatmaps (Trinity)

The Trinity package includes scripts that facilitate the extracting of differentially expressed (DE) transcripts (by FDR and/or log2 fold change) and generating heatmaps and plots for visualization.

Execution

Below command will extract DE transcripts that are at least 4-fold DE at a significance of <= 0.001 in any of the pairwise sample comparisons.

Locate Analysis/DifferentialExpression/analyze_diff_expr.pl in the trinityrnaseq-2.2.0 distribution, and run

cd ~/LSLNGS2015/Trinity/edgeR_trans

PATH_TO_TRINITY/Analysis/DifferentialExpression/analyze_diff_expr.pl \
--matrix ../Trinity_trans.TMM.EXPR.matrix -P 1e-3 -C 2

Option

--matrix location of the matrix.normalized.FPKM file

-P p-value cutoff for FDR (default: 0.001)

-C min abs(log2 fold change) (default: 2, i.e. meaning 2^(2) or 4-fold).

--output prefix for output file prefix for output file (default: diffExpr.P${Pvalue}_C${C})

Output

ls -la diffExpr.*

The R script diffExpr.P1e-3_C2.matrix.R is generated to include all the R commands used.

We use head to examine the first 10 lines of the file containing the DE transcripts at the specified thresholds

head diffExpr.P1e-3_C2.matrix

Use WinSCP to download the PDF file and view it on your PC or laptop.

The diffExpr.P1e-3_C2.matrix.log2.centered.genes_vs_samples_heatmap.pdf shows the clustering result of transcripts along the vertical axis and samples along the horizontal axis.

genes vs. samples heatmap

The diffExpr.P1e-3_C2.matrix.log2.sample_cor_matrix.pdf shows the pairwise sample correlation matrix.

sample correlation matrix
PreviousPerform DE analysis (Trinity)NextVisualization

Last updated

Was this helpful?