These are chat archives for dereneaton/ipyrad

3rd
Apr 2018
Rebecca Tarvin
@frogsicles_twitter
Apr 03 2018 03:02
Hi there, Could you clarify the positioning of the trim_loci parameters? If the locus is AAGGTTTGGTTTACC, what gives TTTGGTT? Is it 0,4,0,4?
Isaac Overcast
@isaacovercast
Apr 03 2018 19:39
@frogsicles_twitter For single end reads only the first 2 numbers are used. To generate the behavior you asked about you'd use trim_loci = 4,-4
This trims 4 bases off 5' and 4 bases off 3' (-4 in this case means 4bp from the end).
Rebecca Tarvin
@frogsicles_twitter
Apr 03 2018 19:52
@isaacovercast What about for loci from PE data?
Isaac Overcast
@isaacovercast
Apr 03 2018 20:24
unmerged PE loci will look like this: AGCGTGCGAGAGTnnnnAGTGCGCGAGCCGA. The first two numbers of the trim_loci param will trim the first part of the fragment and the second two numbers will trim the second part. In both cases the number specify trimming distance in bp from the 5' end of their particular part of the fragment (negative numbers count from the 3' end). For merged PE reads it will only use the first 2 numbers, since there is only one fragment to trim. Does that help?
Let me know if this doesn't make sense.
Rebecca Tarvin
@frogsicles_twitter
Apr 03 2018 21:11
@isaacovercast I am trying to remove the enzyme cut sites to speed up step 3. Now I see the actual way to do this is through trim_reads, not trim_loci (my bad). But out of curiosity, technically the merged locus will have R2 flipped compared to the reads we see in the R2 fastq, is that right? But the only way to trim the reads from both ends is with trim_reads, not trim_loci, then.
Isaac Overcast
@isaacovercast
Apr 03 2018 22:43
@frogsicles_twitter Yes, in the final assembly R2 will be reverse complemented so it is has the same orientation as R1. You can trim the reads both raw unassembled reads and assembled loci from both ends. It's tricky because PE merged loci are treated slightly different than unmerged loci, but it should still work.