Hello,
I sent some samples for bulk RNA sequencing at a core. They prepped all the samples together but then they run the samples in 2 sequencing runs. The sequencing machine they use has 2 sides so they run the samples in side A and side B at the same time. Now I have 2 sets of fastq files- one from side A and one from side B. Can I merge the fastq files by using "cat" or should I be worried about batch effects? For sample A I got 15 million reads from side A and 16 million from side B and I was aiming for 30 million reads.
Thank you
2 answers
Generally batch effects are caused by everything prior to the actual sequencing (sampling, extraction, library prep etc); so personally I would concatenate them.
The sample will come from the same prep but run again in another run. Would that also be ok to concatenate?
Yes. Unless the library is different or sequencing is done with a different chemistry, it should be fine to concatenate the data.
You could also process them in parallel (for alignments) and then merge the BAM files before counting to "speed" things up.
I think it would probably be fine to concatenate them, but you could also check whether the sequencing machine side effect is substantial or negligible using PCA.
I have one case where the same library was sequenced on different instruments (NovaSeq 6000 and NovaSeq X), and when I visualized PCA, the samples clustered very tightly.
but you could also check whether the sequencing machine side effect is substantial or negligible using PCA.
If there was such a effect then more than likely something went wrong with the sequencing (to be clear there should be no effect for a particular library)..
Log in to answer this question.
-
Can I merge fastq files from different runs?
written by bioinfo •
-
Can I compare kallisto counts from samples with different amount of reads?
written by bioinfo •
-
Batch effect consideration (re-seq the same sample twice)
written by jkim •
-
Can these RNAseq runs be combined?
written by mdrnao
-
Merged BAM has more reads than sum of parts
written by James Reeve
-
Technical replicates of same library samples
written by Rosmaninho •
-
Best practice for merging across lanes
written by dave •
-
Problem with memory requirements for SPades
written by Tospohh •
-
Pooling fastq reads from different samples in metabarcoding
written by lvogel •
-
merge .fastq reads of different lengths and align
written by User6891
