rustynuc

Tool to calculate the likelihood of 8-oxoG damage based on alignment characteristics.

Install

Binary

Precompiled binaries are provided below:

| | | | :-----------------------------: | :-------------------------------------: | | TAR | TAR | | ZIP | ZIP |

Cargo

If you have cargo installed or have installed RUSTUP, you can install directly from github bash cargo install --git https://github.com/bjohnnyd/rustynuc

Build

To compile from source rustup is required and can be obtained HERE. After installing rustup download the release archive file and build:

bash git clone https://github.com/bjohnnyd/rustynuc.git && cd rustynuc && cargo build --release

All releases and associated binaries and archives are accessible under Releases.

Usage

bash ./rustynuc -h

``` rustynuc 0.2.0

USAGE: rustynuc [FLAGS] [OPTIONS]

FLAGS: -a, --all Whether to process and print information for every position in the BAM file -h, --help Prints help information -n, --no-qval Skip calculating qvalue -p, --pseudocount Whether to use pseudocounts (increments all counts by 1) when calculating statistics -V, --version Prints version information -v, --verbosity Determines verbosity of the processing, can be specified multiple times -vvv -w, --with-track-line Include track line (for correct visualization with IGV)

OPTIONS: --alpha FDR threshold [default: 0.2] -b, --bcf BCF/VCF for variants called on the supplied alignment file --bed A BED file to restrict analysis to specific regions --fisher-sig Significance threshold for Fisher's test [default: 0.05] --max-depth Maximum pileup depth to use [default: 1000] -m, --min-reads Minimum number of aligned reads in ff or fr orientation for a position to be considered [default: 4] -q, --quality Minimum base quality to consider [default: 20] -r, --reference Optional reference which will be used to determine sequence context and type of change -t, --threads Number of threads [default: 4]

ARGS: Alignments to investigate for possible 8-oxoG damage ```

Output

The default output (if not --bcf/-b is provided) is a BED file with the following info:

1. Chromosome 2. Start 3. End 4. Name (format is `<chromosome>_<start>_<end>` or if reference is provided `<chromosome>_<base>_<start>_<end>` 5. -log10 of p-value (p-value is the smallest of the A/C and G/T ) 6. Strand 7. Depth 8. Adenine FF:FR counts 9. Cytosine FF:FR counts 10. Guanine FF:FR counts 11. Thymine FF:FR counts 12. A/C two-sided p-value Fisher's Exact Test 13. G/T two-sided p-value Fisher's Exact Test (14). Sequnce Context (if reference provided) 14/15. adj. pvalue 15/16. Significant at set FDR value (1 if yes, 0 if not)

To get only positions with p-value below 0.05:

bash rustynuc -r tests/input/ref.fa.gz tests/alignments/oxog.bam | awk '$12 < 0.05 || $13 < 0.05' | gzip > sig.bed.gz

If a VCF/BCF is provided the output will be in VCF format. Multiple summaries are provided in the VCF file:

| TYPE | ID | Description | | :-----------------------------: | :-----------------------------: | :-------------------------------------: | | FILTER | OxoG | OxoG two-sided p-value < 0.05 | | FILTER | InsufficientCount | Insufficient number of reads aligning in the FF or FR orientation | | INFO | OXODEPTH | OxoG Pileup Depth | | INFO | ADENINEFFFR | Adenine counts in FF and FR orientations | | INFO | CYTOSINEFFFR | Cytosine counts in FF and FR orientations | | INFO | GUANINEFFFR | Guanine counts at FF and FR orientations | | INFO | THYMINEFFFR | Thymine counts at FF and FR orientations | | INFO | ACPVAL | A/C two-sided p-value | | INFO | GTPVAL | G/C two-sided p-value | | INFO | FFFRAF | Alternate frequency calculations on the FF and FR (2 values for each alternate allele) | | INFO | OXOCONTEXT | 3mer reference sequence context |

AF_FF_FR can be used to filter based on AF on the FF or FR orientations.

For each alternate allele, there are two AF provided so for example to filter the first alternate positions AF_FF_FR[0] and AF_FF_FR[1] can be used. The command below will filter using the AF on FF/FR and also FILTER=="PASS" ensures only position with p-val < 0.05 are returned.

bash FILTERCMD='TYPE =="snp" && AF > 0.04 && FILTER=="PASS" && (FF_FR_AF=="." || (FF_FR_AF[0] >= 0.04 && FF_FR_AF[1] >= 0.04))' rustynuc --pseudocounts -r tests/input/ref.fa.gz --b tests/input/oxog.vcf.gz tests/alignments/oxog.bam | bcftools filter -Oz -i "$FILTERCMD" > nonoxog.vcf.gz

Authors

• Johnny Debebe

License

The MIT License (MIT). Please see License File for more information.

Notes

• DEPTH is the key determinant of power in discerning 8-oxoG
• in cases where depth is not high the AF_FF_FR alternate frequency filter is a better option
• fisher's exact is affected heavily by 0 counts so pseudocounts can be used
• fisher's exact test is fairly conservative so might underestimate the true numbers
• FDR will be heavily dependent on %GC of the genome, size of the genome, whether a reference was provided, a VCF is provided or the test was restricted to specific regions.

Crates to Credit

Implemented using the rust-htslib and niffler crates.

Citing

If used in published research, a citation is appreciated:

Debebe, Bisrat J: Quick analysis of pileups for likely 8-oxoG locations. (2020). doi:10.5281/zenodo.4157557