Awesome

Introduction

HTSbox is a fork of early HTSlib. It is a collection of small experimental tools manipulating HTS-related files. While some of these tools are already part of the official SAMtools package, others are for niche use cases in my own work, so are maintained by myself. Please keep in mind that HTSbox is NOT the official repository. It lags far behind HTSlib in terms of features, activity, clarity and robustness. If you are looking for a high-quality library and related tools, the SAMtools organization repositories are the right place.

Usage by examples

Summary pileup:
```
htsbox pileup -f ref.fa sorted1.bam sorted2.bam
```
This gives the number of observed alleles, including InDels, in each input BAM. Additional options can be applied for filtering:
```
htsbox pileup -f ref.fa -Q20 -q30 -l90 sorted.bam
```
This filters alignments shorter than 90bp and with mapping quality lower than 20 and filters bases with quality lower than 20.
Naive variant calling:
```
htsbox pileup -f ref.fa -Q20 -q30 -cs3 sorted.bam
```
The output in VCF gives positions with ALT alleles appearing 3 or more times. Note that variant calling in this way is very crude. It is usually not recommended to use this for whole-genome and exon variant calling from short reads. Nonetheless, `pileup' is a proper tool to call variants from contigs, in particular unitigs produced by fermi:
```
htsbox pileup -cuf ref.fa unitig.bam
```
Option -u enables multiple settings.

Generate consensus FASTA:

htsbox pileup -f ref.fa -Q20 -q30 -Fs3 sorted.bam

Pairwise alignment summary:
```
htsbox samview -p aln.bam
```
In the output, each line gives QName, QLen, QStart, QEnd, Strand, RName, RLen, RStart, REnd, PerBaseDivergence, MapQ and semicolon-delimited misc information.
Count alignment break points (mainly used to evaluate misassemblies):
```
htsbox abreak name-srt.sam.gz
```
or call structural variantions:
```
htsbox abreak -u name-srt.sam.gz
```
Reduce quality resolution with Illumina binning:
```
htsbox qualbin -t2 -bm7 in.bam
```
This is the only command so far that explores multi-threading.
Evaluate empirical base quality, similar to MAQ's mapcheck:
```
htsbox mapchk sorted.bam ref.fa
```
The output is a bit complicated. Let me know if you are interested (I will see if it is worth documenting the output).
Generate per-base read depth:
```
htsbox depth -p0 sorted.bam
```
The last two columns give the total read depth and depth of reads mapped with high mapQ (threshold defaults to 20). When -p takes a value between 0 and 1, it will use the CODOC strategy to find windows with relatively stable read depth. The output can be much smaller as it loses some information.