Bioinformatics, computational biology, and data science updates from the field. Occasional posts on programming.

TL;DR:
 
 Codeium offers a free Copilot-like experience in Positron. You can install it from the Open VSX registry directly within the extensions pane in Positron.

Stephen Turner

Positron doesn't have GitHub Copilot, but Codeium and others are good alternatives.

AI code completion in Positron

I'm still catching up on papers from my late 2024 backlog. This week’s recap highlights autonomous microbial sensors for detecting TNT in soil, genome size estimation from long reads, STABIX for indexing and compressing GWAS summary statistics, and Clair3-RNA for deep learning-based small variant calling on long-read RNA-seq data.

Autonomous microbial sensors for detecting TNT in soil, genome size estimation from long reads, indexing + compressing GWAS summary stats, Clair3-RNA for small variant calling on long-read RNA-seq...

Weekly Recap (Jan 2025 part 2)

Happy New Year! I’m still catching up on papers from my late 2024 backlog.

Directed evolution by a protein language model, AI learning to run transcript assemblers, a review/introduction to pangenomics, Alphafold2 refinement for protein design, metagenomic binning...

Weekly Recap (Jan 2025, part 1)

I am in the middle of writing a review / perspectives paper. One that I’m confident will be exciting once we get it published. Some sections of the review cover subject matter at the outer periphery of my expertise.

Listen now | Inciteful with the Zotero plugin is the perfect combo for filling in gaps in your knowledge of the literature in a particular field. BioloGPT and Semantic Scholar are useful too.

Inciteful+Zotero to find relevant literature

If you use ChatGPT, Claude, or even some local model through Ollama or HuggingFace Assistants, you’ll know that the chat interface makes it challenging to feed in an entire repo like a Python or R package, because functions, tests, etc. can be scattered across many files throughout a repo.

The open source repo2txt web app will take a GitHub repo URL, displays the directory structure, lets you choose which files to include, and provides a single plain text file you can feed into any LLM.

Turn a GitHub repo into a single text file for LLM-friendly input

It’s a short week here in the US. As I reflect on the tools that shape modern bioinformatics and data science it’s striking to see how far we’ve come in the 20 years I’ve been in this field. Today’s ecosystem is rich with tools that make our work faster, better, enjoyable, and increasingly accessible.

Data science and bioinformatics tech I'm thankful for in 2024: tidyverse, RStudio, Positron, Bluesky, blogs, Quarto, bioRxiv, LLMs for code, Ollama, Seqera Containers, StackOverflow, ...

Tech I'm thankful for (2024)

This week’s recap highlights compendium of human gene functions derived from evolutionary modelling from the Gene Ontology Consortium, an AI reasoning model applied to rare disease diagnosis, an agentic AI for scRNA-seq data exploration, and applying FAIR principles to scientific workflows.

Comprehensive evolutionary modeling of human gene function, an AI reasoning model for rare disease diagnosis, an AI agent for scRNA-seq data exploration, FAIR scientific workflows, ...

Weekly Recap (May 2025, part 2)

This week’s recap highlights PhyloSketch for interactively drawing and manipulating phylogenies, Uncalled4 for nanopore DNA and RNA modification detection, Severus for SV calling from long reads, CREsted for modeling synthetic cell type-specific enhancers, and a review on transformers and genome language models.

PhyloSketch for drawing phylogenies, Uncalled4 for DNA/RNA modification detection, Severus for long read SV calling, CREsted for synthetic enhancer modeling, a review on genome language models, ...

Weekly Recap (June 2025, part 1)

It’s been a few weeks since I wrote a recap about what I’m reading. It’s been difficult watching helplessly as the institutions and financial infrastructure underpinning my profession are being systematically and irreversibly dismantled, with brilliant scientists I know personally having their careers destroyed and lives upturned.

Predicting RNA-seq coverage from DNA sequence with Borzoi, doing it faster with Flashzoi, phylogenetic inference from structure, prioritization in GWAS vs rare variants, LLMs for patient interactions

Weekly Recap (Feb 2025, part 2)

This week’s recap highlights Verkko2 for T2T genome assembly, the GPN-MSA DNA language model trained on multispecies alignments for variant effect prediction that outperforms other methods like CADD, ESM-1b, phyloP, phastCons, nucleotide transformer, and HyenaDNA, fast orthology inference with FastOMA, a foundation model of transcription across cell types, and engineering CRISPR-Cas PAM sites using deep learning.

T2T assembly, DNA language models for variant effect analysis, ortholog inference at scale, a foundation model of transcription across cell types, engineering PAM recognition, &amp; a few great reviews...

Weekly Recap (Feb 2025, part 3)

Last year I wrote a post describing an R package I put together that fetches recent bioRxiv preprints from a given subject and summarizes them in a couple of sentences using a local LLM running through Ollama:  That tool has a limitation in that it’s using the bioRxiv RSS feed to pull recent paper titles and abstracts, and the RSS feeds currently only provide the 30 most recent preprints in each subject area.

Collect metadata and publication details for &gt;200k preprints over a 10 year period, investigate trends, and scrape full text for sentiment analysis

Exploring the bioRxiv API with R, httr2, rvest, tidytext, and Datawrapper

I’ve been blogging about genetics, statistics, computational biology, data science, and science in general for over 15 years. I published my first blog post on Getting Genetics Done in 2009, and started this blog last year after taking a few years off.

 Science blogging has significantly contributed to my personal and professional growth as a scientist.

Writing takes time and effort — time I could have spent elsewhere.

Writing a technical blog improves your writing, forces you to learn new things, helps others and yourself, and helps your career. This post: why blog, what to blog about, how to start, and using AI.

Technical blogging for growth and learning

This post is inspired by the Bluesky Network Analyzer made by @theo.io.

I’m encouraging everyone I know online to join the scientific community on Bluesky. In that post I link to several starter packs — lists of accounts posting about a topic that you can follow individually or all at once to start filling out your network. I started following accounts of people I knew from X and from a few starter packs I came across.

Find people followed by the people you follow, but who you don't follow, using R and the atrrr package

Expand your Bluesky network with R

This week's recap highlights a new pipeline for metagenome quality assessment and taxonomic annotation (MAGFlow &amp;

Nextflow for metagenomics QA &amp; taxonomic classification, SNP/SV benchmarking, SNV calling with nanopore data, assembly polishing with transformers, Nextflow for variant simulation &amp; evaluation, ...

Weekly Recap (Nov 2024, part 1)

I just returned from a week in Barcelona where I attended the Nextflow Summit and nf-core hackathon, and I can hardly contain my excitement for the near term future of bioinformatics, computational biology, and open science in general.

Highlights from the 2024 Nextflow Summit and nf-core hackathon in Barcelona

Nextflow Summit Barcelona 2024

Yesterday I wrote about base R vs. dplyr vs. duckdb for a simple summary analysis. In that post I simulated 100 million rows of a dataset and wrote to disk as CSV. I then benchmarked how long it took to read in and compute a simple grouped mean. One thing I

 didn’t

do here was separate the time it took to read data into memory (for base R and dplyr) versus computing the actual summary.

Parquet is interoperable between Python and R, fast to read+write, works well with databases, and stores complex data types (e.g., tibble listcols). Use it instead of CSV. Many pros, few (no?) cons.

Use nanoparquet instead of readr/CSV

TL;DR
 
 : For a very simple analysis (means by group on 100M rows), duckdb was 125x faster than base R, and 28x faster than readr+dplyr, without having to read data from disk into memory. The duckplyr package wraps DuckDB's analytical query processing techniques in a dplyr-compatible API. Learn more at duckdb.org/docs/api/r and duckplyr.tidyverse.org.

I wanted to see for myself what the fuss was about with DuckDB.

Using DuckDB in R to analyze 100 million rows of data in 3 seconds

DuckDB vs dplyr vs base R

This week’s recap highlights a new multispecies codon optimization method, personalized pangenome references with vg, a commentary on the wild west of spike-in normalization, a new pipeline for comprehensive and scalable polygenic scoring across ancestrally diverse populations, a paper showing deep learning / transformer-based methods don’t outperform simple linear models for predicting gene expression after genetic perturbations, and finally, a

What I'm reading this week on multispecies codon optimization, personalized pangenome references, spike-in normalization, automating polygenic scoring, transformers vs linear models, bactoneurons, ...

Weekly Recap (Oct 2024, part 1)

Cite:

Stephen Turner. “Learning in Public.”

 Paired Ends

(2024). DOI: https://doi.org/10.59350/xwgsf-nj906. I wrote my first public blog post in 2009. I started Getting Genetics Done to share what I was learning at the end of my PhD/postdoc through my first few years as faculty.

A short essay on writing about what you're learning

Learning in Public

This week’s recap highlights a new tool from Wei Shen and Zamin Iqbal for efficient sequence alignment against millions of prokaryotic genomes (LexicMap), a new tool from Heng Li for efficiently constructing and querying a sequence index at scale, an R/Bioconductor package for detecting and correcting DNA contamination in RNA-seq data, a method for dating gene age using synteny, how AlphaFold predictions for some types of conformations are

What I'm reading this week in alignment and indexing millions of genomes at scale, gene age and family analysis and visualization, synthetic enhancers and cell fate programming, single cell omics, ...

Weekly Recap (Sep 2024, part 3)

This week’s recap highlights perspectives in de-extinction and patent law, systematic benchmarking of scATAC-seq methods, a 91 gigabase (!) animal genome, structural variant genotyping with long reads, cell type-specific enhancer prediction, and a perspective piece in AI in biosecurity.

What I'm reading this week in de-extinction and patent law, ATAC-seq benchmarking, large animal genomes, SV genotyping with long reads, cell type-specific enhancer prediction, AI in biosecurity, ...

Weekly Recap (Sep 2024, part 1)

It’s been a big week in the genomics+bioinformatics space. This post expands on a few of the recent papers I posted in a Twitter thread. I highlight a few in the deep dive at the top, then link to a few other papers of note later on. Subscribe (free!) to Paired Ends to get summaries like this delivered to your e-mail as soon as I write them.

What I'm reading this week in SV analysis with long reads, spatial transcriptomics, gene prediction and annotation, synthetic biology, AI in genomics &amp;

What I'm reading (Aug 2024, part 1)

TL;DR

I wrote an R package that summarizes recent bioRxiv preprints using a locally running LLM via Ollama+ollamar, and produces a summary HTML report from a parameterized RMarkdown template.

The biorecap package summarizes bioRxiv preprints using a local LLM. I used two LLMs to help me write the package, and two other LLMs to help me write a paper about the package, published on arXiv.

biorecap: an R package for summarizing bioRxiv preprints with a local LLM

This post expands on a few of the papers I posted in this Twitter thread. I highlight a few in the deep dive at the top, then link to a few other papers of note later on. Subscribe to Paired Ends to get summaries like this in your e-mail as soon as I write them.

What I’m reading this week in synthetic biology, population genetics, AI in genomics/bioinformatics, microbial genomics, and more

What I'm reading (July 2024)

This week’s recap highlights a new method for gene-level alignment of single-cell trajectories, an R package for integrating gene and protein identifiers across biological sequence databases, characterization of SVs across humans and apes, universal prediction of cellular phenotypes, a method to quantify cell state heritability versus plasticity and infer cell state transition with single cell data, and a new AI-driven, natural language-oriented

Gene-level alignment of scRNA-seq trajectories, integrating identifiers across databases, SVs in humans and apes, universal prediction of cellular phenotypes, and a bioinformatics analysis chatbot

Weekly Recap (Oct 2024, part 2)

Anyone reading this newsletter has surely used the frontier models like ChatGPT, Claude, and Gemini. I’ve written a few posts about using

 local

models but haven’t really talked much about the tools I use to directly interact with these models. Those previous posts interact with local models using tools like ellmer in R or my own biorecap package which interacts with a locally running Ollama server.

Running local LLMs with Ollama, LMStudio, Open WebUI, and others for private and secure chat, vibe coding, image analysis, and "chat with your documents" using RAG

GUIs for Local LLMs with RAG

This week’s recap highlights a new nf-core workflow for multi-omics trait association studies, a new tool for linking genotype to phenotype (G2P) by directly sequencing alleles from CRISPR base editing experiments, the SplitsTree app for interactive analysis and visualization using phylogenetic trees and networks, mapping cellular interactions from spatially resolved transcriptomics data, a study of marine microbial diversity and bioprospecting

Regulatory element mapping, phylogenetic tree analysis+viz, mapping cellular interactions with scRNA-seq, bioprospecting, transformers in single-cell omics, nf-core/omicsgenetraitassociation, ...

Weekly Recap (Sep 2024, part 4)

This week’s recap highlights an AI agent for automated multi-omic analysis (AutoBA), rapid species-level metagenome profiling and containment (sylph), a review on genome-wide association analysis beyond SNPs, private information leakage from scRNA-seq count matrices, and a method to “unlearn” viral knowledge in protein language models as a means to develop safe PLM-based variant effect analysis (PROEDIT).  Others that caught my attention include

What I'm reading in agentic multiomic analysis, metagenome profiling, GWAS beyond SNPs, private information leakage from scRNA-seq, unlearning viral knowledge for safe protein language models, ...

Weekly Recap (Nov 2024, part 2)

This week’s recap highlights a new Nextflow workflow for calculating polygenic scores with adjustments for genetic ancestry, a paper demonstrating that whole exome plus imputation on more samples is more powerful than whole genome sequencing for finding more trait associated variants, a new deep-learning-based splice site predictor that improves spliced alignments, a new method for accurate community profiling of large metagenomic datasets, and

Nextflow workflow for polygenic scoring, WES+imputation &gt; WGS, DL method for splice site prediction, metagenomics profiling, and a new NIST/GIAB matched tumor-normal with 13 sequencing modalities

Weekly Recap (Oct 2024, part 3)

This week’s recap highlights a Nextflow pipeline for eQTL detection, an end-to-end pipeline for spatial transcriptomics (visium) data analysis, a method for identification of perturbed cell types in single cell RNA-seq data, a method for guide assignment in single-cell CRISPR screens, a tool for on-target/off-target analysis of gene editing outcomes, and “digital microbes” for collaborative team science on emerging microbes.

A new eQTL Nextflow workflow, spatial transcriptomics / scRNA-seq, guide assignment in CRISPR screens, on-/off-target analysis after CRISPR editing experiments, digital microbes, paleogenomics, ...

Weekly Recap (Sep 2024, part 5)

This week’s recap highlights Google/Deepmind’s new AlphaProteo tool for protein design, tools for protein structure alignment and analysis, biases in polygenic risk scores due to overlap and kinship, highly variable gene selection in single cell RNA-seq, and reconstruction of a 4.2 billion year old last universal common ancestor of life on Earth (spoiler alert: CRISPR-Cas is &gt;4B years old!).  Others that caught my attention include a new

What I'm reading this week in protein design with AlphaProteo, biases in polygenic risk scores, scRNA-seq HVG selection, reconstruction of a 4.2 billion year old last universal common ancestor, ...

Weekly Recap (Sep 2024, part 2)

Here are a few papers that caught my attention recently. I summarize a few in the deep dive at the top, then link to a few other papers of note later on. Subscribe to Paired Ends to get periodic summaries like this delivered to your e-mail.

What I'm reading this week in genomics, bioinformatics, cancer biology, epigenome editing, biosecurity, synthetic biology, single-cell omics, population genetics, microbial genomics, ...

What I'm reading (Aug 2024, part 2)

The Baader–Meinhof phenomenon (aka the frequency illusion) is the name for that thing that happens when you buy a new car, and suddenly you notice that same model car everywhere you drive.

Something a little different this week: book recommendations! De-extinction is a common theme across a few of the fiction and nonfiction books I read this year.

What I'm reading: de-extinction edition

This post expands on a few of the papers I posted in this Twitter thread. I highlight a few in the deep dive at the top, then link to a few other papers of note later on. Subscribe to Paired Ends to get summaries like this delivered to your e-mail as soon as I write them. Subscribe now You might remember me from Getting Genetics Done where I blogged about genetics, statistics, and bioinformatics from 2008-2017.

What I’m reading this week in synthetic biology, population genetics, AI in genomics/bioinformatics, pangenomes, and more

What I'm reading (July 2024, part 2)

This is Paired Ends — your bioinformatics, computational biology, and data science updates from the field.

Coming soon

What I'm reading this week in single cell omics, spatial transcriptomics, kinship and selection inference in archaeogenomics, population and medical genetics, ...

What I'm reading (Aug 2024, part 4)

There was a time in late 2023 to early 2024 when I and probably many others in the R community felt like R was falling woefully behind Python in tooling for development using AI and LLMs. This is no longer the case. The R community, and Posit in particular, have been on an absolute tear bringing new packages online to take advantage of all the capabilities that LLMs provide.

Python isn't the only game in town anymore: R can interact with local and cloud LLM APIs, inspect and modify your local R environment and files, implement RAG, computer vision, NLP, evals, &amp;

The Modern R Stack for Production AI

Llama 3.1 405B is the first open-source LLM on par with frontier models GPT-4o and Claude 3.5 Sonnet. I’ve been running the 70B model locally for a while now using Ollama + Open WebUI, but you’re not going to run the 405B model on your MacBook.

How to create and deploy a free GPT4-class chatbot on HuggingFace Assistants for any GitHub repo, using an R package as an example, in less than 60 seconds

Create a free Llama 3.1 405B-powered chatbot on a GitHub repo in &lt;1 min

This week’s recap highlights the WorkflowHub registry for computational workflows, building a virtual cell with AI, a review on bioinformatics methods for prioritizing causal genetic variants in candidate regions, a benchmarking study showing deep learning methods are best for variant calling in bacterial nanopore sequencing, and a new ML model from researchers at Genentech for predicting cell-type- and condition-specific gene expression across

WorkflowHub computational workflow registry, building a virtual cell with AI, bioinformatics for prioritizing causal variants, DL methods for nanopore variant calling, predicting gene expression, ...

Weekly Recap (Dec 2024, part 1)

I’ve been using the llama3.1:70b model just released by Meta using Ollama running on my MacBook Pro. Ollama makes it easy to talk to a locally running LLM in the terminal (ollama run llama3.1:70b) or via a familiar GUI with the open-webui Docker container. Here I’ll demonstrate using the ollamar package on CRAN to talk to an LLM running locally on my Mac.

Using R to prompt llama3.1:70b running on my laptop with Ollama + ollamar to tell me what's interesting about a set of genes, and to summarize recent bioRxiv preprints

Use R to prompt a local LLM with ollamar

I recently stumbled across Phil Ewels’s ~18 minute nf-core/bytesize talk on Excalidraw: For years I’ve been using draw.io for making flowcharts and diagrams for documentation, papers, presentations, and for general brainstorming and communication with my team, clients, and collaborators.1 Excalidraw (excalidraw.com) looks like an attractive alternative.

Excalidraw might be my new go-to open-source tool over draw.io for creating and sharing workflow diagrams. It's simple, fast, aesthetically pleasing, and has amazing VS Code integration.

Excalidraw: create and share workflow diagrams with end-to-end encryption

Google has a new experimental1 tool called Illuminate (

 illuminate.google.com

) that takes a link to a preprint2 and creates a podcast discussing the paper. When I tested this with a few preprints, the podcasts it generated are about 6-8 minutes long, featuring a male and female voice discussing the key points of the paper in a conversational style. There are some obvious shortcomings.

Listen now | Google's new Illuminate tool takes a PDF from arXiv and generates a short podcast discussion on the paper's key points

Illuminate preprints with an AI-generated podcast discussion

This is an experiment... ignore this post

Paired Ends Podcast

This post is about the R package development experience with Positron, the new IDE from Posit based on VS Code. This is not a tutorial on R package development in general — there are great resources for that elsewhere. Read on.Subscribe to Paired Ends to get future posts like this delivered to your e-mail.

 RStudio, VS Code, and Positron

Back in 2011 I wrote a blog post about a relatively new IDE for R called RStudio.

First impressions using Positron for R (and Python!) package development

R package development in Positron

Last month I published a paper and an R package for summarizing preprints from bioRxiv using a local LLM. I wrote about it here: Llama 3.2 was just released today (announcement). The biggest news is the addition of a multimodal vision model, but I was intrigued by the reasonably good performance of the tiny 3B text model. I used this as an excuse to update the biorecap R package.

The biorecap package can now summarize bioRxiv and medRxiv preprints, and uses the llama-3.2-3b model by default.

Llama 3.2 summaries of bioRxiv and medRxiv preprints with biorecap

VP (Pete) Nagraj
 
 is a long time friend, colleague, and collaborator, and is the author of this post. Pete and I have co-authored over a dozen publications, and have taught several graduate courses in data science together. Pete leads the health security analytics / infectious disease modeling and forecasting group at Signature Science, where we started this work together last year.

PLANES provides a set of methods for evaluating the plausibility of epidemiological signals and forecasts. The PLANES methods are available in the rplanes R package and Shiny app.

PLANES: Plausibility Analysis of Epidemiological Signals

One of my previous employers was a Google Cloud partner, which gave me full and free access to all of Google Cloud’s certification programs, where I took the Professional Cloud Architect and Professional Data Engineer programs. It shouldn’t surprise anyone that with Google leaning hard into GenAI that they have new certification programs and learning paths, like this Generative AI Leader certification.

Free short courses on GenAI, LLMs, responsible AI, image generation, encoder-decoder architectures, BERT, attention, transformers, ...

Free AI courses from Google

In the spirit of learning in public, I wanted an excuse to dive into Quarto to learn more about publishing formats beyond simple PDF and HTML documents. If you’re not familiar, Quarto (quarto.org) is the successor to RMarkdown, the next-generation scientific publishing system that works natively with Python, R, and OJS. If you already have RMarkdown you probably don’t have to do

 anything

to it to get it to render with Quarto.

Turn a collection of RMarkdown documents into a book in &lt;1 hour using Quarto

Writing a book with Quarto

I get asked a lot how I have the time to read all these papers and articles I post about here. I don’t. Not all of them at least. I

 listen

to many of them. Lately I’ve been using an app called

 ElevenReader

, made by the popular text-to-speech service ElevenLabs. It’s free for iOS and Android.

ElevenReader (free for iOS/Android) will read articles and PDFs in realistic voices. Example clips here with Burt Reynolds, Maya Angelou, and Richard Feynman reading recent articles and preprints.

Listen to papers, PDFs, and articles with ElevenReader

This week’s recap highlights FLAMES for prioritizing genes at trait-associated GWAS hits, integrating protein language models and an automatic biofoundry for enhanced protein evolution, benchmarking DNA sequence models for causal regulatory variant prediction, and the doubletrouble R/Bioconductor package for identifying and classifying gene and genome duplications.

Effector genes at GWAS hits, protein evolution, regulatory variant prediction, gene/genome duplication analysis, genomic foundation models, Nextflow pipelines for annotation and polygenic scores, ...

Weekly Recap (April 2025, part 2)

Here we are most of the way through March and I’m just getting around to my first “weekly” recap. It’s been a busy month — I’m writing a few papers of my own which I’ll share here when they’re published, and I took some much needed R&amp;R in Portugal where I traded my stack of research papers for some escapist sci-fi. But I’m back now and making my way through a deep backlog.

Heritable polygenic editing, self-supervised learning in single-cell genomics, predictive modeling with tabular foundation models, ESM3, pangenome graphs in biodiversity genomics, ...

Weekly Recap (March 2025, part 1)

R 4.5.0 was released last week, and Bioconductor 3.21 came a few days later. You can read the R release notes here and the Bioconductor 3.21 announcement here.

Faster package installation, import only the functions you want with use(), built-in Palmer penguins data, grep values shortcut, and lots of new bioinformatics packages in Bioconductor

R 4.5.0 and Bioconductor 3.21

This week’s recap highlights Evo2 for variant effect analysis and genome design, a preprint showing that pretraining doesn’t necessarily increase performance on genomic foundation models, a new R package ggalign for making complex biological data visualizations with ggplot2, and an ancestral reconstruction method for ancient DNA. I also highlight a few reviews in biodiversity genomics.

Evo2, genomic foundationless models, the ggalign R package, ancestral reconstruction for ancient DNA, biodiversity genomics reviews, structural variation in mammalian genomes, pangenomics, ...

Weekly Recap (April 2025, part 1)

In a previous post I demonstrated how to set up a local LLM that you can run through either a command line interface (Ollama) or a graphical user interface (Open WebUI and others), and quickly demonstrated how to “chat with your documents” with a local model using LMStudio. In that previous post I simply attached a few documents to a one-off chat.

Demonstration using Mistral and Nomic Embed with Ollama and Open WebUI

Build a local RAG application with Open WebUI to chat with your Zotero library

This week’s recap highlights ESCARGOT, an AI agent for biomedical knowledge graphs and reasoning, CASTER for direct species tree inference from whole-genome alignments, the scGPT-spatial foundation model for spatial transcriptomics, the BioChatter platform for biomedical research applications with LLMs, moscot for mapping cells through time and space, and two reviews: one on epigenetic clocks and another on structural variation in the human

AI agents for biomedical reasoning, species tree inference, a spatial transcriptomics foundation model, biomedical research with LLMs, mapping cells through time and space, epigenetic clocks, SVs, ...

Weekly Recap (March 2025, part 2)

This week’s recap highlights pangenome graph construction with nf-core/pangenome, building pangenome graphs with PGGB, benchmarking algorithms for single-cell multi-omics prediction and integration, RNA foundation models, and a Nextflow pipeline for characterizing B cell receptor repertoires from non-targeted bulk RNA-seq data.

Pangenome graph construction with nf-core/pangenome, pangenome graphs, single cell multi-omics, RNA foundation models, Nextflow for B-cell receptor sequencing from nontargeted bulk RNA-seq, ...

Weekly Recap (Nov 2024, part 3)

This week’s recap highlights protein design with RoseTTAFold, surveillance with wastewater sequencing, T2T human genomes, Vitessce for visualization of multimodal spatial single-cell data, and Taxometer for taxonomic classification of metagenomics contigs.

Protein design with RoseTTAFold, wastewater sequencing T2T human genomes, single-cell spatial transcriptomics, taxonomic classification, ...

Weekly Recap (Oct 2024, part 4)

This is part 2 of a series on uv. Other posts in this series:


 uv, part 1: running scripts and tools


 This post


 uv, part 3: Python in R with reticulate


 Coming soon…

Last year I wrote a post on creating a Python command line application with Click using a cookiecutter template, building with setuptools and publishing with twine.

Using uv to create, manage, build, and publish a Python package to PyPI

uv, part 2: building and publishing packages

We’re 6 weeks into the new year and I’m

 still

catching up on papers from my late 2024 backlog.

Neuropsychiatric PGS don't predict professional categories, pangenome graph-based GWAS, building and using single cell atlases, sketching sketching in genomic epidemiology, multimodal protein model...

Weekly Recap (Feb 2025, part 1)

This is part 1 of a series on uv. Other posts in this series:


 This post


 uv, part 2: building and publishing packages


 uv, part 3: Python in R with reticulate


 Coming soon…

Lately I’ve heard a lot great things about uv, an extremely fast Python package and project manager, written in Rust.

Over the past year I’ve heard a lot great things about uv, an extremely fast Python package and project manager, written in Rust.

uv, part 1: running scripts and tools

I'm still catching up on papers from my late 2024 backlog. This week’s recap highlights a browser application for visualizing pathogen dispersal, a DNA language model evaluation benchmark on regulatory DNA, regularized ensemble polygenic risk prediction with GWAS summary statistics, multimodal analysis of RNA-seq data for complex trait genetics, and a deep dive on blastp’s E-value.

Visualizing pathogen dispersal, DNA language model benchmarks for regulatory DNA, polygenic risk prediction with GWAS summary stats, multimodal analysis of RNA-seq data, blastp's E-value, ...

Weekly Recap (Jan 2025 part 3)

Something a little different for this week’s recap. I’ve been thinking a lot lately about the practice of data science education in this era of widely available (and really good!) LLMs for code. Commentary at the top based on my own data science teaching experience, with a deep dive into a few recent papers below.

Listen now | Balancing growth with guidance: teaching data science in the age of AI assistants

AI in data science education

This week’s recap highlights the Evo model for sequence modeling and design, biomedical discovery with AI agents, improving bioinformatics software quality through teamwork, a new tool from Brent Pedersen and Aaron Quinlan (vcfexpress) for filtering and formatting VCFs with Lua expressions, a new paper about the NHGRI-EBI GWAS Catalog, and a review paper on designing and engineering synthetic genomes.

The Evo model for sequence modeling and design, AI agents in biomedical research, improving bioinfo software with teamwork, VCF manipulation, GWAS Catalog, designing+engineering synthetic genomes...

Weekly Recap (Dec 2024, part 3)

The majority of developers use LLMs to help write code, present company included. When I’m working in languages I know well, they're fantastic at handling the grunt work: generating boilerplate, suggesting completions, and writing tedious tests and documentation.

I used Claude to build a Chrome extension to add a post-to-Bluesky button on bioRxiv. I've never written Javascript or a Chrome extension before. It took 5 minutes.

Write code in unfamiliar territory with AI

OpenAI introduced the ability to create custom GPTs back in November 2023. I wanted to try to create one of these, and in the spirit of learning in public this post describes how I made it. But first, what does it do?Gene Info Custom GPT

 Gene Info custom GPT

The Gene Info custom GPT takes a list of human gene symbols as input.

Gene Info: a custom GPT that takes a list of gene symbols and provides summary information, gene ontology terms, and provides contextual information like pathway or disease involvement.

Gene Info Custom GPT

Last week I posted about a web app that turns a GitHub repo into a single text file for LLM-friendly input. This is great for capturing LLM-friendly text from a GitHub repo, but what about any other arbitrary website or PDF? I was catching up on Simon Willison’s newsletter reading about an app he made with Claude artifacts that uses the Jina Reader API to generate Markdown from a website. You don’t need to use the API to do this.

Add r.jina.ai/ in front of any URL to turn the entire page into Markdown for LLM-friendly input. Works on PDFs too.

Turn any webpage into markdown for LLM-friendly input

Background


 Bluesky, atrrr, local LLMs

I’ve written a few posts lately about Bluesky — first, Bluesky for Science, about Bluesky as a home for Science Twitter expats after the mass eXodus, another on using the atrrr package to expand your Bluesky network. I’ve also spent some time looking at R packages to provide an interface to Ollama.

Summary, translation, and sentiment analysis of Bluesky posts on a topic using atrrr, tidyverse/ellmer, and mlverse/mall with Claude and other open models, and gistr to put the summary on GitHub.

Bluesky conversation analysis with local and frontier LLMs with R/Tidyverse

I had good intentions to give NaNoWriMo a try this year but didn’t get very far. Instead I gave OpenAI’s Creative Writing Coach GPT a try for a (very) short story I had in mind, inspired by my frustration trying to access closed-access research articles for a review article I’m preparing.

A parable of scientific publishing, open access, and preprints