LengthIns=0, "StartTE/EndTE" are "NA"

[WeijiaSu]

Hi @adamewing I am trying to use TLDR with Drosophila ONT sequences, I only tested it with one particular TE (LTR) as the reference TE insertion, I got the result .txt file with no error reported. However, I the "LengthIns" column of all (~300) detected insertions is 0, and the "StartTE/EndTE" are "NA". I have a few known insertion sites that should be around 6000-7000 bp and they are also labeled as LengthIns=0. Thanks for your help. Weijia

[adamewing]

Are these filtered insertions (i.e. something other than "PASS" in the Filter column)?

[WeijiaSu]

Thanks for your reply. The Filter column shows "LeftFlankSize,RightFlankSize,UnmapCoverNA,NoTEAlignment,NonRemappable,ShortIns"; I used similar command line as the run_test.sh: tldr -b test.sort.bam -e MyTestTE.fa -r MyGenome.fasta --color_consensus -p 16 --max_te_len 100000 --detail_output Thanks for your help. Weijia

[adamewing]

Thanks. That will be why there is no TE length as those filters indicate there was no good TE alignment for that insertion (do all insertions have this set of filters or do some "PASS")? A few questions:

• how was the .bam file aligned? I've mostly tested with minimap2 .bams.
• what does the sequence ID from MyTestTE.fa look like?
• Do you see filtered insertion calls at the expected TE insertion positions?

[WeijiaSu]

Thank you. All insertions have the same filter, no entry labeled as "PASS". For the questions

1. I used minimap2.bam
2. The sequence ID is >HMS-Beagle; (only 1 TE in there)
3. I did see a few output entries that are overlapping with the expected insertions. I also manually check the raw reads from a reported insertion by the *.detail.out files, these 36 supported reads all aligned to the TE, some even aligned more than 6000 bp.

[adamewing]

I think I managed to replicate this - it's a documentation bug: I haven't explained what's expected in terms of the te ref file fed to -e. The .fasta header should include a superfamily (e.g. L1) and a subfamily (e.g. L1Ta) seperated by a ":" (see ref/teref.human.fa for examples). So try changing ">HMS-Beagle" to ">TE:HMS-Beagle" and let me know what happens.

[adamewing]

Added a check for formatting on .fasta file passed to -e/--elts in a38312f, it should refuse to use a te ref file without the right formatting and explain what that should look like now.

[WeijiaSu]

Hi, I re-ran the pipeline as you suggested, looks like it worked properly. Thank you for your help! Weijia

[adamewing]

Great, will close this then.