Which Sample to use for analysis when I have same sample from different lanes?

Question

[Nikhil]

I have one pair end sample with 3 Different Lanes.

Lane 1:- V300110856_L01_A01_1.fq.gz V300110856_L01_A01_2.fq.gz

Lane 2:- V300110856_L01_A02_1.fq.gz V300110856_L01_A02_2.fq.gz

Lane 3:- V300110856_L01_A03_1.fq.gz V300110856_L01_A03_2.fq.gz

So should I merge the lane data or Can i go ahead analyzing data from one Lane? They are prepared from same library.

[antonioggsousa]

Hi,

You should merge the 3 lanes by orientation, i.e., forward and reverse, and use the merged/combined files for analysis.

In bash you can merge the files by doing:

cat V300110856_L01_A01_1.fq.gz  V300110856_L01_A02_1.fq.gz V300110856_L01_A03_1.fq.gz > V300110856_1.fq.gz
cat V300110856_L01_A01_2.fq.gz V300110856_L01_A02_2.fq.gz V300110856_L01_A03_2.fq.gz > V300110856_2.fq.gz

I hope this helps,

António

Answer 1

[Pierre Lindenbaum]

map each pair or read in parallel, set the read-group according to the lane ( https://gatk.broadinstitute.org/hc/en-us/articles/360035890671-Read-groups eg. PU ).

then merge the 3 bams.

[Nikhil]

Okay. So it is necessary that we merge the lanes data?

[Pierre Lindenbaum]

yes if you need to process the data per sample (eg VCF calling)

[Nikhil]

Okay, Thank you. What will happen if I process the sample just using one lane data?

[Pierre Lindenbaum]

less depth, less quality

[Nikhil]

Ohh, Thank you for the information. Its still fine if I merge at fastq level, right?

[a.bibek52]

Hi! Pierre, I am having the same query regarding merging reads of multiple lanes together. However, I am not being able to find any documented sources till date, neither any of the literature tell about whether to merge the reads or perform the downstream to individual PE reads from individual lanes. Can you please help me with the documentation, if available?

Thank you.