finds.dev

bigdatagenomics/adam

ADAM is a genomics analysis platform with specialized file formats built using Apache Avro, Apache Spark, and Apache Parquet. Apache 2 licensed.

View on GitHub

★ 1,060ScalaApache-2.0updated Mar 2026

// the take

ADAM is a Scala library and CLI that puts genomics file formats (BAM, VCF, BED/GFF3/GTF) into Avro-defined schemas on Parquet, so you can process reads, variants, and features as Spark Datasets instead of shelling out to single-purpose tools. It's for bioinformatics teams with sequencing data at a scale where cluster-distributed processing actually pays off, not for someone running a handful of exomes.

Doing everything in-memory on Spark avoids the usual genomics pipeline pattern of writing intermediate files between every tool, which is a real bottleneck at scale. The Parquet+Avro schema approach gives actual schema evolution and lets you query genomic data with Spark SQL instead of parsing SAM/VCF text by hand. There's a real ecosystem built on its APIs (Avocado for variant calling, Cannoli for wrapping BWA/FreeBayes, Mango for visualization), which suggests the core abstractions are solid enough for others to build on. Installable via Conda, Homebrew, and Docker, not just a Maven build from source.

You need a Spark cluster (3.2.0+ specifically) to get the benefit, which is real infrastructure and ops overhead compared to single-node tools like GATK or samtools that most bioinformatics shops already run. Migrating existing BAM/VCF pipelines onto ADAM means actually converting your data into its Parquet-backed format, not just pointing it at existing files. The two canonical papers describing its design are from 2013 and 2015 — the architecture predates the broader field's shift toward WDL/Nextflow plus elastic single-node containers, so the pool of teams who want cluster-resident Spark for genomics specifically is narrower than when this was built. It's also a research-lab-maintained project (bigdatagenomics org), so support and release cadence depend on academic funding cycles rather than a company with a commercial stake in uptime.

We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →

// at a glance

Stars
1,060
Forks
312
Language
Scala
License
Apache-2.0
Last commit
17 Mar 2026

// nearby

If adam isn't quite it.

All repos →