De novo transcriptome assembly using Trinity

Question

[1234anjalianjali1234]

Hi guys,

I have a total of 12 samples [6 infected sample (number- 7,8,9,10,11 and 12) of different time-point with their 6 control samples (number- 1,2,3,4,5 and 6)] from RNAseq, paired end (Illumina; GAIIx).

My Sample was distributed in 8 Lanes. Details:

LANE      SAMPLE
LANE_1: 7,8,9,10          [Control samples]
LANE_2: 1,2,3,4,5,6       [Infected samples]
LANE_3: 11,12,7,8         [Control samples]
LANE_4: PhiX control lane
LANE_5: 9,10,11,12        [Control samples]
LANE_6: 1,2,3,4           [Infected samples]
LANE_7: 5,6,1,2           [Infected samples]
LANE_8: 3,4,5,6           [Infected samples]

The data provided to me was in bcl format, which i have converted them into fastq format using bcltofastq software. Now, I have 16 fastq files, both forward and reverse:

Right Reads    Left Reads
LANE_1_R2.fq   LANE_1_R1.fq
LANE_2_R2.fq   LANE_2_R1.fq
LANE_3_R2.fq   LANE_3_R1.fq
LANE_4_R2.fq   LANE_4_R1.fq
LANE_5_R2.fq   LANE_5_R1.fq
LANE_6_R2.fq   LANE_6_R1.fq
LANE_7_R2.fq   LANE_7_R1.fq
LANE_8_R2.fq   LANE_8_R1.fq

My problems:

1. For the denovo assembly do i need to feed all the R1 reads as left and R2 reads as right in TRINITY and do I need to remove PhiX Lane.

2. How to align reads from each sample to the reference because i dont have seperate fastq file of each sample.

Many thanks in advance!

Answer 1

[JC]

1. In general yes, you can pass all the transcriptome to Trinity, check the "normalize" parameters to reduce the large assembly memory issue. Alternatively, you can assemble samples per treatment (I generally see better assembly because isoforms are less per sample) and then merge your assemblies.

2. The Illumina Fastq IDs should indicate the index (https://en.wikipedia.org/wiki/FASTQ_format#Illumina_sequence_identifiers), and you need to know which index belongs to each library to demultiplex your files.

[1234anjalianjali1234]

Thankyou JC,

I will try this.