I ran some mm9 data through the HISAT2/StringTie/Ballgown RNA-seq pipeline recently and went through the listed protocol HISAT2/StringTie/Ballgown Protocol. When I used Ballgown, I was not able to identify any novel transcript isoforms as compared with Table 3 from data analyzed in the linked paper. The first thing I tried was to use no filter, but that did not work. I am wondering if there was an issue in how I fed the data into Ballgown in terms of my experimental conditions?

My experiment consists of two experimental levels. The first being a genotype of SMAD4 KO or WT. The second being a treatment of either TGF-beta, BMP, or Ctrl. In total, I have six experimental groups.

Right now I have the following lines of code when trying to analyze the differential expression of the genes and transcripts.

DE by transcript

results_transcripts <- stattest(bg_mm9, feature='transcript', covariate='genotype', adjustvars=c('treatment'), getFC=TRUE, meas='FPKM')

DE by gene

results_genes <- stattest(bg_mm9, feature='gene', covariate='genotype', adjustvars=c('treatment'), getFC=TRUE, meas='FPKM')

I appreciate any advice or guidance. Please let me know if you think there might be any issues besides the experimental grouping/leveling that is preventing me from getting novel transcript isoforms as found in Table 3 of the linked Nature protocol paper. Thank you so much.