Fastq files with very large numbers of reads could take a very long time to run through the workflow. To address this, we are currently removing any samples with > 1500 megabases of sequence data with this augur filter parameter:
--query "(mbases > 180 & mbases < 1500 & (country != 'Uncalculated'))"
Instead, we could address this issue by randomly downsampling sequence reads for very large fastq files. We could potentially downsample these files to 1 gigabase, since that should be more than enough reads for high genotyping quality.
This downsampling would need to take into account the fact that samples differ in read length and whether they have single end or paired end reads.
This was discussed on Slack.
Fastq files with very large numbers of reads could take a very long time to run through the workflow. To address this, we are currently removing any samples with > 1500 megabases of sequence data with this
augur filterparameter:Instead, we could address this issue by randomly downsampling sequence reads for very large fastq files. We could potentially downsample these files to 1 gigabase, since that should be more than enough reads for high genotyping quality.
This downsampling would need to take into account the fact that samples differ in read length and whether they have single end or paired end reads.
This was discussed on Slack.