Checking correlation on raw fastq data without aligning

Question

[agata]

Maybe a stupid general question - is it any tool available that would allow for checking the correlation between raw fastq files of biological replicas?

Do I have to always first align reads to reference and then use e.g DeepTools/DeSeq2 packages to obtain information about the correlation of the samples?

[LauferVA]

could you define what you mean by correlation in this context?