Is it right to combine multiple runs downloaded from SRA?

Question

[jeni]

Hi!

This may be a naive question, but I am downloading some data from SRA. I am downloading all runs used in a specific paper which come from the same cell line and were sequenced following the same library preparation protocol and in the same sequencer.

The fastq that I get from a unique run is shorter than those that I normally use, so I am not sure if I should combine all the fastqs and analyze then the complete fastqs (obviously maintaining fastq1 and fastq2).

Thanks!

[ATpoint]

Can you share an accession number?

[jeni]

For example, from this bioproject: PRJNA268125

[ATpoint]

If these are lange replicates then you can merge them. For HiC it is not unusual to require a sample being split over many lanes. In this bioproject you have to check for methods texts or metadata that indicate what these samples are and if these are sequencing/lane replicates.

[lieven.sterck]

I would personally only join them if they are noted as technical replicates, never join biological ones