Skip to content
View dnbaker's full-sized avatar

Block or report dnbaker

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this userโ€™s behavior. Learn more about reporting abuse.

Report abuse
dnbaker/README.md

Hi, I'm Daniel ๐Ÿ‘‹

Software Engineer at Roche. Previously, I was a Senior Scientist at Pacific Biosciences (PacBio) after earning my PhD at Johns Hopkins University in the department of Computer Science. Before that I was a Bioinformatics Scientist at ARUP Laboratories, where I worked on cell-free circulating tumor DNA (ctDNA) analysis and clinical genomics after my training in Physics [BS] and Biophysics/Computational Biology [MS]. I've worked with biological data (sequence, molecular modeling, metabolomics, transcriptomics, metagenomics), telecommunications data, as well as graph algorithms, machine learning, and numerical optimization.

๐Ÿ”ญ I've worked on similarity search, and clustering, and indexing for large-scale biological data, simd/gpu-accelerated and randomized algorithms. Most recently, I've been developing methods for human genetics, including long RNA-seq, VNTRs, and haplotype phasing.

๐Ÿ˜„ Pronouns: He/Him/His

A quick tour of my interests

  1. Practical randomized algorithms

This ranges from libraries providing sketch data structures and coresets, as well as projects using random projections and DCI.

My work on coresets and clustering is primarily part of the minicore project, with the aims of providing a standard utility for coreset construction and weighted clustering, especially for exponential family models and shortest-paths metrics.

  1. Computational Biology

The bonsai project provides methods for metagenomic analysis, along with k-mer encoding/decoding and I/O, while the Dashing performs scalable sketching and comparison of sequence data.

BMFtools performs molecular demultiplication over sequencing barcoded data, reducing error rates while eliminating redundant information. Designed for ctDNA, this method can reduce error rates by orders of magnitude, allowing confident detection of very rare events.

scavenger has rust implementations using tch-rs for VAEs for count-based data, applied to single-cell transcriptomics.

I also co-developed pbfusion, a fast tool for characterizing transcriptional abnormalities.

  1. General C++

Most of my projects fall into this category, serving as tools I can reuse in various projects.

Some of my favorites:

  • vec provides type-generic abstractions over x86-64 vectorization, making it easy to write fast, portable code.
  • kspp is an RAII-based variant of kstring from klib with extra niceties making appending printf-style formatting easy.
  • aesctr provides STL-style random number generators built on fast aes-ctr and wyhash
  • circularqueue provides a range-based circular queue container that uses power-of-two sizes

Pinned Loading

  1. sketch sketch Public

    C++ Implementations of sketch data structures with SIMD Parallelism, including Python bindings

    C++ 154 13

  2. frp frp Public

    FRP: Fast Random Projections

    C++ 43 6

  3. ARUP-NGS/BMFtools ARUP-NGS/BMFtools Public

    Barcoded Molecular Families

    C++ 22 8

  4. minicore minicore Public

    Fast and memory-efficient clustering + coreset construction, including fast distance kernels for Bregman and f-divergences.

    C++ 33 6

  5. dashing2 dashing2 Public

    Dashing 2 is a fast toolkit for k-mer and minimizer encoding, sketching, comparison, and indexing.

    C++ 62 7

  6. bioseq bioseq Public

    Tokenizers and Machine Learning Models for biological sequence data

    C++ 24 4