input hybrid nanopore and illumina data

[eesiribloom]

Im wondwering if its possible to incorporate both illumina and nanopore data in to xTea? The graphical abstract makes it look like this is possible but the hybrid approach only mentions x-linked reads with illumina. Or should I run separately on nanopore and illumina data and then compare results?

[simoncchu]

It works in a hybrid mode (taking advantage of both short and long reads). But it is not fully exported as an automatic option. Here are the detailed steps to run in the hybrid mode:

1. Run only on the Illumina short reads data. You can only run the first "clip" step -f 1, and you'll get the output from this step in a name like: candidate_list_from_clip.txt. Or, you can just the whole pipeline line (-f 5907), which you'll generate all the intermediate files.
2. Run the long-read mode on the nanopore data, and in the assembly step (-A step), set the Illumina output as an input (--i2 option). An example script for this step is like:

PREFIX=/n/data1/HG002/HG002_ccs/combine_Illumina/
REF=/n/data1/GRCh38/GCA_000001405.15_GRCh38_no_alt_analysis_set.fna
XTEA_PATH=/n/data1/projects/xTEA_long_reads/
BAM_LIST=${PREFIX}"bam_list.txt"
TMP=${PREFIX}"tmp/"
REP_LIB=/n/data1/XTEA/rep_lib_annotation/rep_lib_annotation/
############
ILLUMINA_SRC=/n/data1/GIAB/Illumina_hg38_NIST_60X/candidate_list_from_clip.txt
############
python ${XTEA_PATH}"l_main.py" -A -b ${BAM_LIST} -r ${REF} -p ${TMP} -i ${PREFIX}"candidate_list_from_clip.txt" --i2 ${ILLUMINA_10X_SRC}  -o ${PREFIX}"all_ins_seqs_with_illumina.fa" -n 8

In this example, we set the intermediate output candidate_list_from_clip.txt from the Illumina short-reads data as the input, and you can also change to intermediate output from other steps. The basic idea is that if the sequencing depth for the nanopore data is low, and it will miss some candidates at its first step, and by integrate the intermediate output from short reads, we may rescue these when run the assembly step.

[eesiribloom]

thanks I will give this a try and flag if I have any issues

[eesiribloom]

Can I just clarify...
From running xtea with illumina short reads I get 4 candidate list files:

SVA/candidate_list_from_clip.txt
HERV/candidate_list_from_clip.txt
L1/candidate_list_from_clip.txt
Alu/candidate_list_from_clip.txt

Which of these is the input in to the long read pipeline? or do I need to merge them together?

[simoncchu]

Sorry for the late response. For certain TE type, it's better to feed in the TE-specific "candidate_list_from_clip.txt".