Expression analysis by Trinity

Question

[satomuraatsushi]

Hi guys, I need your kind support:)

I am doing de novo RNA-seq analysis by using Trinity. At first I combined left reads and right leads, which were obtained from three different conditions.

cat right1.fastq right2.fastq right3.fastq > R1.fastq

cat left1.fastq left2.fastq left3.fastq > R2.fastq

And I assembled these reads by Trinity to produce assembled Fasta file (Trinity.fasta).

Trinity --seqType fq --right R1.fastq --left R2.fastq --SS_lib_type RF -min_contig_length 500--CPU 8 --max_memory 16G

Finally, I mapped the individual reads to the created Trinity.fasta. The below code shows the mapping of only right1.fastq and left1.fastq, but I did the same thing for the remaining fastq files.

align_and_estimate_abundance.pl --right1.fastq --left left1.fastq --est_method RSEM --aln_method bowtie2 --trinity_mode --prep_reference --output_dir ./output_files --transcripts Trinity.fasta

However, looking at the number of hits in the three RSEM.isoforms.results, there were many Trinity reads that were not hit under any conditions. Since the Trinity reads were constructed by the fastq data, I think it is very strange that no hits were created.

Are there any possible reasons for this phenomenon?

[Brian Bushnell]

In this kind of scenario, more information is always helpful. Please describe what kind of data you are using, the goal of your experiment, your exact command lines at every step, etc. Without that it's hard to give a useful response.

[satomuraatsushi]

Thank you for your quick reply:) mRNA was extracted from a kind of Mite. The goal of this experiment is to evaluate the mRNA expression levels without mapping to the known genome sequence. I used Miseq reagent kitV3 (600 cycle: 300*2 paired end). The first step in the analysis was trimming using cutapdat and TrimmomaticPE to remove polyA sequences and low-quality reads.

cutadapt -a AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 1_S1_L001_R1_001.fastq > 1_R1_trimed.fastq

TrimmomaticPE 1_R1_trimed.fastq 1_R2_trimed.fastq 1_S1_L001_R1_001_paired.fastq 1_S1_L001_R1_001_unpaired.fastq 1_S1_L001_R2_001_paired.fastq 1_S1_L001_R2_001_unpaired.fastq TRAILING:32 MINLEN:100 SLIDINGWINDOW:4:30

I confirmed the better quality of the trimmed reads by fastqc. I did the same procedures for the remaining data (2_S2_L001_R1_001.fastq, 2_S2_L001_R2_001.fastq , 3_S3_L001_R1_001.fastq, and 3_S3_L001_R2_001.fastq). I combined the thee fastq files.

cat 1_S1_L001_R1_001_paired.fastq 2_S2_L001_R1_001_paired.fastq 2_S2_L001_R1_001_paired.fastq > R1.fastq

cat 1_S1_L001_R2_001_paired.fastq 2_S2_L001_R2_001_paired.fastq 2_S2_L001_R2_001_paired.fastq> R2.fastq

Assembling and the subsequent steps were the same as above.