FAQ

FAQ

Does z-fastq repair malformed FASTQ?

No. It reports the first problem and stops; it never skips records or guesses a fix. See Validation.

Does it load whole files into memory?

No. Every command streams, and exact-count sampling keeps only a small entry per chosen record. See Limits.

Does the .gz name decide gzip decoding?

No, the first two bytes do. See Input and output.

Why does check print nothing on success?

So it can be a plain pass or fail step in a script: the exit status is the answer.

Which quality encoding does it read?

Phred+33 only, ASCII 33 to 126.

Can I sample an exact count from a pipe?

No, exact counts read the file twice. Use --fraction on a pipe. See Sampling.

Are paired mates ever sampled separately?

Never. Both mates are kept or neither.

Does it write gzip?

Only process --gzip. Pipe the other commands' output to gzip.

Can an output file be overwritten?

No. deinterleave and process create new files and fail on an existing path.

Does it support wrapped FASTQ or FASTA?

Not today. See Limits.

How fast is it, and how much memory does it use?

On one core, faster than the other tools that do a whole job, with a few MiB of memory, and its memory does not grow with the input. Some tools that do part of a job, or use many threads, are faster in places. See Performance.

Is z-fastq a replacement for fastp, FastQC, or SeqKit?

No. It follows fastp's trimming rules and FastQC's QC definitions for the parts it covers, and does much less than any of them. See Choosing a FASTQ tool.