Merging Mapped Reads from Technical Replicates

[Fernas]

Dear All,

I am working to analyze (Illumina 2.0 RNASeq paired-end reads) of length 100bp. these reads belong to 10 biological replicates (different samples of the same species) and each sample was sequenced twice (2 technical replicates). So, in total I have the reads from 20 libraries.

My purpose is to study the differences between biological replicates (samples) and I want to get rid of technical replicates. The option of pooling reads from technical replicates is not desired because if an artifact read occur in one replicate will not be detected.

Consequently, I mapped each library (of the 20 libraries) using tophat 2.0.

Now, I want to do the following:
For each sample, I want to go through all the mapped reads (in accepted_hits.bam files) in both replicates and select reads the common reads in both libraries. The purpose of this step is to remove the effect of artifact reads and to select reads that are verified in 2 technical replicates for each samples. Any idea how to do that?

[Fernas]

Can I run (samtools mpileup) and from the outputs select minimum read depth at each position?

[shi]

Dear All,

I am working to analyze (Illumina 2.0 RNASeq paired-end reads) of length 100bp. these reads belong to 10 biological replicates (different samples of the same species) and each sample was sequenced twice (2 technical replicates). So, in total I have the reads from 20 libraries.

My purpose is to study the differences between biological replicates (samples) and I want to get rid of technical replicates. The option of pooling reads from technical replicates is not desired because if an artifact read occur in one replicate will not be detected.

Consequently, I mapped each library (of the 20 libraries) using tophat 2.0.

Now, I want to do the following:
For each sample, I want to go through all the mapped reads (in accepted_hits.bam files) in both replicates and select reads the common reads in both libraries. The purpose of this step is to remove the effect of artifact reads and to select reads that are verified in 2 technical replicates for each samples. Any idea how to do that?

I dont think you should do this. Due to the randomness in the fragmentation step in sample prep, there wont be a large overlap in read sequences between the technical replicates. You should instead combine them rather than get the intersection of them.