An alignment tool targeting the footprints of HERVs in human genomes
HERVranger is a bioinformatics pipeline for the analysis of the footprints of human endogenous retroviruses (HERVs) from human genomes. The hg38 human genome annotation from the UCSC Genome Browser has only partially annotated some HERVs. Therefore, we reviewed and extracted 3,659 known HERV sequences from both the hg38 annotation and the NCBI RefSeq database, and built these HERV sequences as a HERV reference genome for the STAR aligner. The RNA-Seq data of each tumor and normal sample was first aligned to the hg38 genome. FeatureCounts was used to count the expression of the non-HERV genes in the hg38 annotation. We extracted the unmapped reads and the reads that were mapped to the HERVs annotated in hg38. Then we re-aligned them with the HERV reference genome. We performed quantile normalization for the counts of the HERVs together with the expression of the non-HERV genes. In this way we achieved high confident HERV references and expression levels.
Linux (x86_64-redhat-linux-gn) shell (4.2.46(2))
Python (>=2.7.16)
STAR (2.5.2b preferred)
featureCounts (subread-1.5.0 only)
Only run the jobs on 256GB nodes.
Python packages
os, sys, re, shutil, time, pandas, collections, itertools
library1: a STAR reference library build from HERV-removed Hg38 genome reference downloaded from NCBI RefSeq. Please use [/project/SCCC/Wang_lab/shared/HERV_Ref/STAR].
library2: a STAR reference library build from HERV RNA-Seqs. Please use [/project/SCCC/Wang_lab/shared/HERV_Ref/STAR_HERV_092717].
ref.gtf: a featureCounts reference GTF file with HERV removed. Please use [/project/SCCC/Wang_lab/shared/HERV_Ref/hg38mm10.gtf]
output: a path to build a result folder (named by Prefix) to write result files and intermediate files, total usage could be over 10 GB for one pair of RNA-Seqs.
R1.fastq.gz, R2.fastq.gz: a pair of pair-end RNA-Seqs.
To apply HERVranger to single-strand alignments, please input R1.fastq.gz NA, i.e. replace the R2.fastq.gz with a string 'NA' and keep the space between them.
Path_to_saved_RNA_Seqs: where the pair of RNA-Seqs stores.
Prefix: a user decided sample name.
python Realignment_maint.py /path/to/library1 /path/to/library2 /path/to/ref.gtf /path/output R1.fastq.gz R2.fastq.gz /path/to/saved/RNA_Seqs Prefix
If an index file like the example [samples_example.xlsx] is provided, the [jobupload.py] script may be applied to generate large batches of job scripts. The index file is expected to contain one 'Root_tree' column, one 'Seq_ID' column and one 'Path' column.
1.0.0: First release. (09-28-2019)
1.1.0: Update. Single-strand alignment mode was provided. (10-17-2019)