Combining RNA-Seq FASTQ files for coverage issues

[adrian]

Hi:

As a pilot experiment I ordered RNA-Seq on following plan:

3 different conditions
1 sample per condition.
No replicates.

I had 50M reads coverage for each sample 100bp PE.

Second time, I went with 4 replicates. However I have a problem now with coverage:

50bp PE reads

25M coverage for every replicate sample. As there was some space on flow-cell, seq lab ran same libraries on extra lane.

That extra lane also gave 25M coverage.

I was asked to combine FASTq files from two lanes into one FASTQ file suggesting it will make up to 50M reads. It does not make sense to me.

Is that a common practice? what experts suggest here?

Thanks
Adrian

[dpryan]

That's fine, presuming that the libraries run on the two lanes were the same (i.e., a separate library wasn't made for the additional lane). Combining technical replicates is normally OK.

[Jeremy]

The problem with a large difference in read count between samples is that the high read count sample will find low expression genes and the low read count sample will not find those genes which can result in false positives.