I got a fastq files from core and they loaded same library on two different lanes. I read the posts of people that deal with this problem and they recommend to merge these file after converting to bam. My question is next, why I cannot merge these files using tophat into accepted_hits.bam by using these files like technical replicates? What will be the difference between these two methods. I plan to use cufflinks for the analysis of the DE. Thank you for the help.
I assume you mean two different lanes on the sequencer rather than lines.
You can either merge the fastq files or merge the bam files, which should lead to the same final outcome. Using those separately as technical replicate probably has probably limited added value.
Note: Tophat-Cufflinks is no longer the recommended RNA-seq processing pipeline. A new protocol is described here: Transcript-level expression analysis of RNA-seq experiments with HISAT, StringTie and Ballgown, but there are many more alternatives.