Empty results

[jsitarka]

Hi team,

I am using your program in my master thesis development on Arabidopsis genome and I always get empty both .ref.bed and nonref.bed files. Is there any desired format of short reads to get a valid result?

I run command: conda activate ngs_te_mapper2 python3 /path/to/source/code/ngs_te_mapper2/sourceCode/ngs_te_mapper2.py -o output -f SRR8397878_1.fastq -r TAIR10_chr_all.fasta -l RepeatMaskerPlants_nospace.fasta

Reads were downloaded from: https://www.ebi.ac.uk/ena/browser/view/SRR8397878

Reference fasta was downloaded from: https://www.arabidopsis.org/download/index-auto.jsp?dir=%2Fdownload_files%2FGenes%2FTAIR10_genome_release%2FTAIR10_chromosome_files

I used library using in RepeatMasker (but I had to remove spaces and special characters like \,? etc. from id description - because there were some problems during intermediates directories creation e.g. in case there was the line ">Gypsy-17 LBS-I LTR Gypsy" corresponding folder name was only "Gypsy-17" but the program expected full name)

The library has a structure like this:

Gypsy-17_LBS-I_LTR_Gypsy aaggtggacactgtgggaaccaacagcaacctggccggcgtaacagcaga ... BEL-154_AA-LTR_LTR_Pao tgtctacgaccaacaaaacctacttatccctcattactctactggtgcaa ... Copia-1_DYa-I_LTR_Copia ataggttatgggcccaggagtagtaaagactttaataattgtgtgtgatc ...

Message from log file: 04/23/2021 23:31:16: INFO: CMD: ../ngs_te_mapper2/sourceCode/ngs_te_mapper2.py -o output -f SRR8397878_1.fastq -r TAIR10_chr_all.fasta -l RepeatMaskerPlants_nospace.fasta 04/23/2021 23:31:16: INFO: Parsing input files... 04/24/2021 09:52:42: INFO: Start alignment... 04/24/2021 11:49:47: INFO: Alignment finished in 1 hours 57 minutes 4 seconds 04/24/2021 11:49:47: INFO: Detecting insertions... 04/24/2021 14:42:50: INFO: Insertion candidate search finished in 2 hours 53 minutes 2 seconds 04/24/2021 14:43:07: INFO: ngs_te_mapper finished in 15 hours 11 minutes 49 seconds 04/24/2021 14:43:07: INFO: Number of reference TEs: 0 04/24/2021 14:43:07: INFO: Number of non-reference TEs: 0

Am I doing something wrong? Thanks a lot for your advice :)

[shunhuahan]

• @jsitarka Thanks for your interest in testing ngs_te_mapper2 on Arabidopsis genome. We haven't yet tested our program on Arabidopsis dataset but we are happy to get you feedback on whether ngs_te_mapper2 can perform reasonably good on your dataset.
• To fully reproduce your zero prediction results, could you send me your transformed TE library so we can start a replication analysis and figure out what's causing the issue? You can send it over to shhan@uga.edu.
• Also, you might want to try out McClintock, which is an easy to use meta-pipeline for identifying TE insertions using multiple detection methods (including ngs_te_mapper2). You could read details and follow instructions on this page https://github.com/bergmanlab/mcclintock. Our group is currently maintaining this software so feel free to ask questions related to McClintock under its issue page and we will get back to you.

Best, Shunhua

[shunhuahan]

Hi @jsitarka,

• Sorry for the late reply! I’ve got a chance to take a look at your input files and finished a test run. I was able to reproduce the zero prediction issues you reported.
• There are two issues contributing to zero prediction results. One is a bug in the source code that would cause issues when the contig names in the reference genome don't contain chr substring. I fixed the bug in the latest commit (https://github.com/bergmanlab/ngs_te_mapper2/commit/f7281d3261cabe871fea4ad05d305391f5abd2cc).
• Another issue is from your TE library. ngs_te_mapper2, and most other TE detection methods, expect TE consensus sequences (separated by family). I believe one of the TE library you sent me is from https://www.arabidopsis.org/download/index-auto.jsp?dir=%2Fdownload_files%2FGenes%2FTAIR10_genome_release%2FTAIR10_transposable_elements, which includes sequences of all TE insertions in the reference genome. Most TE families in the library fasta file are overrepresented, which is not recommended and might cause issues.
• In my testing, I ran the latest ngs_te_mapper2 using one TE family sequence from TAIR10 database as library file. I was able to get non-zero predictions in the final output.
• Let me know if that makes sense to you and if you have other questions.

Shunhua

[shunhuahan]

• To clarify what the input TE library file should include, I updated ngs_te_mapper2 README in https://github.com/bergmanlab/ngs_te_mapper2/commit/d6ed0941cb847771ba5fa17e82f2ef5b54351bd8.
• Closing this issue for now. @jsitarka Feel free to re-open this issue if the latest update doesn't work for you.