What is the relationship between the fast5 and fastq files in terms of number of reads? MinKNOW enables setting the number of reads per both sample types separately.

Fast5 is a format of raw data coming out from the sequencer. Fast5 file is like a bucket collecting reads from the sequencer. Its capacity is defined by the user in MinKNOW during the run setup. Basecaller translates these reads into fastq files whose capacity is defined again by the user during the run setup.
Example 1: Fast5 = 4000 reads, Fastq = 4000 reads. In this example, 1 fast5 will generate 1 fastq file. So if you collect 20 fast5 files per barcode, it will generate 20 fastq files.
Example 2: Fast5 = 4000 reads, Fastq = 500k reads. In this example, multiple fastq5 files will generate one fastq file. So if you collect 125 fast5 files per barcode, it will generate 1 fastq file.
Example 3: Fast5 4000 reads, Fastq = 2000 reads. In this example, 1 fast5 will generate 2 fastq files. So if you collect 20 fast5 files per barcode, it will generate 40 fastq files.