Transcript quantification/Differential gene expression analysis ?

Question

[sunnykevin97]

Hi newbie to bioinformatics research, I performed de novo assembly on ~100 RNA_Seq data sets from one study with different experimental setup. I got the assembled transcripts and removed the redundant transcripts using clustering algorithms (CD-Hit) Do I use non-redundant transcripts for quantification step or else redundant transcripts for transcript quantification step ? please let me know. Once after quantification is performed all the quantified transcripts subjected to the downstream analysis using edgeR to know the differential gene expression.

Thanks

Answer 1

[h.mon]

Use the full set of transcripts, Trinity has scripts to summarize the quantification to a "gene" level analysis. See the wiki Trinity Transcript Quantification.

[sunnykevin97]

Hi, thanks for the useful information. I started analysis on pilot study, I chosen samples 1 - 32.

1-8 --> Odor  --> Experiment1
9-16 --> Model --> Experiment2
17-24 --> Real --> Experiment3
24--32 --> Control --> Experiment4

I had generate gene map and matrix files for all the samples individually, result looks fair enough. The next step is to perform cross normalization for each experiment is it right ? combining gene_trans_map files and feeding quant.sf files as an input to generate a combine abundance_estimates matrix is it right ? Please correct if I'm wrong.

[sunnykevin97]

Hi, I made gene counts matrix for ~32 samples individually, how do I combine them into one matrix as an input for DESeq2. I'm unable to understand tximport, much simpler way.Suggestions! gene.counts.matrix for one sample

GeneID             10_sample
TRINITY_DN0_c0_g1   12717.85
TRINITY_DN100000_c0_g1  93.65
TRINITY_DN100001_c0_g1  110.58
TRINITY_DN100002_c0_g1  80.76
TRINITY_DN100003_c0_g1  386.84
TRINITY_DN100004_c0_g1  137.85
TRINITY_DN100005_c0_g1  52.14
TRINITY_DN100006_c0_g1  87.19
TRINITY_DN100009_c0_g1  27.88