Which Input is recommanded on step re-annotation, all or filtered?

[zhouzhendiao]

In tutorial preprocessing chapter, all_contig_annotations.csv is used as input. However, in load data for analysis, input changed to filtered_contig_annotations.csv. I wonder which input is recommaned, raw one or filtered one? Does dandelion preprocessing pipeline filter the annotations of contig?

[zktuong]

hi @zhouzhendiao - i apologise for the confusion. In practice i think if you are working on mature T/B cells, then filtered_contig_annotations.csv would probably be alright. We started using all_contig_annotations.csv because we were working on the developing datasets in the past couple of papers by Suo et al. I would suspect that you would probably end up with similar recovery with either formats. The trade off for me is about speed vs how thorough you want your analysis - the all_contig file has a lot of contigs that dandelion will have to run through but will process everything but filter_contig gives you a quick pass at the data. Hope this helps

[zhouzhendiao]

OK, I got it. Thanks!