I am using vsearch to perform dereplication on some SampleData[JoinedSequencesWithQuality] that are obtained through ITSxpress. Problem is that vsearch is returning an error stating that some sequences are empty (Found blank or whitespace-only line before '+' in FASTQ file). I actually looked into the sequences in the .qza from ITSexpress and I can find a lot of weird stuff, i.e. empty sequences and super short sequences. A small example here:
@M01168:5:000000000-A7C98:1:1101:10151:9983 1:N:0:2
AACCC
+
JJJJJ
@M01168:5:000000000-A7C98:1:1101:10155:15804 2:N:0:2
CACCAATCAAGCCTGGCTTGGTATTGGGCGACGGGGTGCACCCGCGCCTCAATTTCTCCGGCTGAACGACCACTATCTCAGCGCTGTGATAATCAATTCGCTGTCGAGACGGGTGCTCACGCCGTTAAAGATTTTATACA
+
JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJGJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ
@M01168:5:000000000-A7C98:1:2101:23914:6789 1:N:0:2
@M01168:5:000000000-A7C98:1:1101:10165:3002 1:N:0:2
TCAAC
+
JIJJJ
@M01168:5:000000000-A7C98:1:1101:10181:5305 1:N:0:2
TCAA
+
JJJJ
Now, I understand there was an issue with a previous version of ITSxpress back in 2018 but it looked like with newer releases of the plugin, it was solved. However, here I am. I also found this other post reporting same issue by vsearch but I believe it's not a vsearch issue but an ITSxpress-related issue.
I run the same data with the DADA2 plugin and no problem here (it was also suggested in the same post i already linked). However, I wanted to perform the analysis on the same samples using both OTU and ASV approach but currently I can't go through vsearch due to this issue. I am also considering the possibility that trimming ITS sequences and then dereplicating and clustering with vsearch shouldn't be done but then...why not?
It might be worth noting that I am using qiime2-2021.8 with itsxpress version 1.8.0 and q2-itsxpress version 1.8.0.
EDIT 19th October 2022:
there are some news concerning this post on its github page.