tag:blogger.com,1999:blog-36768584.post3709370786589020966..comments2024-03-03T18:49:34.382-05:00Comments on Omics! Omics!: Does any analytical program really care about the order of paired end files?Keith Robisonhttp://www.blogger.com/profile/04765318239070312590noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-36768584.post-9048159406822229922016-02-19T06:47:53.299-05:002016-02-19T06:47:53.299-05:00While in a theory there shouldn't be much diff...While in a theory there shouldn't be much difference in the quality of the R1 / R2 reads, anybody who had tried processing some MiseQ's 2x250 or 2x300 bp runs (esp from some slightly dodgy reagent batches) would tell you that the difference can be huge due to cluster expansion between cycles and other factors, which cause the the R2 reads to fall in quality much more sooner/deeper. <br /><br />Quite often it is beneficial to add 20-30 cycles to R1 from R2 (if you have single end barcoding).<br /><br />Storage wise - have a common repository with CONSISTENT raw fastq naming schema (can be symlinks) and have a consistent fastq_dir structure for the intermediate fastq processing results within each project.<br /><br />I always call R1/R2 specifically, newer rely on a chance...<br /><br />Usually different analysis like de novo, mapping, have different optimums for best input format/preprocessing.<br /><br />Also not all of them take compressed files, so the use of 2-4TB drives as a local worknodes scratch disks is quite common :-)Marknoreply@blogger.comtag:blogger.com,1999:blog-36768584.post-4969384617065288182016-02-03T18:23:02.556-05:002016-02-03T18:23:02.556-05:00RNAseq aligners and bisulfite-sequencing aligners ...RNAseq aligners and bisulfite-sequencing aligners might (more likely "will" for BS-seq aligners) both break if you swap read #1 and #2.Anonymoushttps://www.blogger.com/profile/00429869551651155257noreply@blogger.comtag:blogger.com,1999:blog-36768584.post-37767145497704751542016-01-23T20:49:46.258-05:002016-01-23T20:49:46.258-05:00Definitely in downstream tools the read # matters,...Definitely in downstream tools the read # matters, say for duplicate detection (certain types), if there are inline barcodes, or strand-specific sequencing. Adapter trimming is another one. There are others. Feed in R1 before R2 and all will be forgiven. icemanhttps://www.blogger.com/profile/07610627797253022010noreply@blogger.comtag:blogger.com,1999:blog-36768584.post-16312980891473171462016-01-22T13:32:28.424-05:002016-01-22T13:32:28.424-05:00Guy: Ugh, yes I tried to use ASCII art in a foolis...Guy: Ugh, yes I tried to use ASCII art in a foolish way. Now to go edit that...Keith Robisonhttps://www.blogger.com/profile/04765318239070312590noreply@blogger.comtag:blogger.com,1999:blog-36768584.post-49260079436736772142016-01-22T11:16:35.201-05:002016-01-22T11:16:35.201-05:00Keith, you wrote:
Particularly important is tracki...Keith, you wrote:<br />Particularly important is tracking what the relationship is between those ends and how they were generated, since paired ends generate reads that face into each other (--> <-- away="" each="" face="" from="" mate="" nbsp="" other="" pairs="" typically="" whereas="">).<br /><br />Is the parenthetical part a typo of some sort (using left and right angle brackets as arrowheads is a real pain in html), or do you just find keeping track of the whole mate pairs/paired ends/f-r/r-f/r-r/f-f kerfuffle as infuriating as I do?Guyhttps://www.blogger.com/profile/02968872267548865219noreply@blogger.com