data layout from SRA is not reliable

[aguang]

The data layout (i.e. paired or single-ended) field from SRA is not reliable. This had led twice now to issues where bioflows expects the data to be paired end and searches for the pair, but it is not, so the workflow errors out. In one case SRX1726841 fastq-dump split the file into 2 by just putting the first half of the reads in 1 file and the second half of the reads in the other, but the read IDs did not actually match since they weren't paired.

One potential solution to this is to write a check on the fastq files to see if the first x read IDs match up. However this would require processing through the fastq files in addition to when FASTQC already runs on them, so this would increase bioflows run time, and may not be desirable behavior.