BioKotlin¶

BioKotlin is a high-performance bioinformatics library that brings the power and speed of compiled programming languages to scripting and big data environments.

It supports nucleotide and protein sequence manipulation, fast sequence IO, k-mer analysis, genomic ranges, and GFF feature trees. Because it runs on the JVM, it interoperates with the wider genomics ecosystem: HTSJDK, GATK, BioJava, and TASSEL.

Get started Browse the tutorials

A quick look¶

Sequences are immutable and type-safe. DNA and RNA share one backing store, and the alphabet is inferred from the sequence unless you name it explicitly.

import biokotlin.seq.*

fun main() {
    //sampleStart
    val dna = NucSeq("GCAGAT")         // DNA inferred from the 'T' amino acid
    println(dna.complement())          // CGTCTA
    println(dna.reverse_complement())  // ATCTGC
    println(dna.transcribe())          // GCAGAU
    println(dna.translate())           // AD
    //sampleEnd
}

Protein sequences are a separate type, so the compiler stops you from, say, adding DNA to a peptide. The AminoAcid enum carries the properties you usually have to look up.

import biokotlin.seq.*

fun main() {
    //sampleStart
    val protein = ProteinSeq("GCAGAT") + ProteinSeq("ARSQRS")
    println(protein) // GCAGATARSQRS
    println("Gly count: ${protein.count(AminoAcid.G)}")

    var mass = 0.0
    for (i in 0 until protein.size()) mass += protein[i].weight
    println("Mass: $mass daltons")
    //sampleEnd
}

Why Kotlin?¶

Kotlin is a high-performance language that runs on the Java Virtual Machine. It is fully interoperable with Java, but its syntax is closer to Python and other functional languages, and it has a number of features designed for scripting and domain-specific languages.

Because Kotlin is compiled, it can be many times faster than scripting languages for the same work - up to two orders of magnitude on the benchmarks we track. BioKotlin also stores DNA with two bits per base pair, which saves four- to eight-fold on RAM with only modest performance losses.

Where BioKotlin can, it copies BioPython's beautifully designed syntax; see the BioPython comparison.

Kotlin for data science¶

With GraalVM making JVM languages and other scripting languages (Python, R) interoperable, Kotlin works with all the most popular data science environments.

• Kotlin in Jupyter supports a notebook environment alongside Python's numpy and R's dplyr and ggplot.
• GraalVM is a polyglot environment supporting FastR, JVM, Python, and C languages.

What's in the library¶

The API reference documents all of them.

Where to go next¶

• Getting started - install BioKotlin in Jupyter, a Kotlin script, or a Gradle project.
• Tutorials - worked examples, generated from runnable notebooks in the repository.
• Contributing - we welcome new contributors.