Carl de Boer

Group Leader - Genome Biology Unit Are you ready to lead groundbreaking research in Genome Biology? Join us at EMBL! We are seeking a motivated scientist to lead an independent research group addressing exciting and original biological...

To all post-docs: The Genome Biology dept ‪@embl.org
has an Independent faculty position. Fantastic place to set up your lab –great package: core funding, fantastic Ph.D. students, cutting edge core facilities & great colleagues. Closing date Sept 19th
embl.wd103.myworkdayjobs.com/en-US/EMBL/j...

GAME: Genomic API for Model Evaluation The rapid expansion of genomics datasets and the application of machine learning has produced sequence-to-activity genomics models with ever-expanding capabilities. However, benchmarking these models ...

Thanks for reading! Please let us know what you think, and support GAME by contributing modules!
Preprint: doi.org/10.1101/2025...
GitHub: github.com/de-Boer-Lab/...

GAME was designed in consultation with many functional genomics and ML researchers. Many enthusiastically contributed their models and datasets in GAME modules. Thanks to all the coauthors and colleagues for your support in developing GAME! 🙏

As GAME builds momentum, more models and benchmarks will be created, and, because they’re all inherently cross-compatible, the easiest way to benchmark your model will be to put it in GAME, snowballing further.📈

6) Better estimates of real-world performance (since tasks and models are cleanly separated and uniformly applied)
7) GAME enables continual benchmarking; we anticipate a yearly “state of the field” to be easily produced. Here's an example for gene expression prediction using existing GAME modules.

5) Getting funding 💰and people for long term maintenance of bioinformatics projects is notoriously challenging😱. Our solution is inherently sustainable in the long term because it is distributed: anyone can add their own GAME modules!🔥

3) GAME communicates over TCP/IP sockets, enabling evaluation of even remote/proprietary models (think ChatGPT)🌐
4) Because the Matcher is modular, we can swap it with a better version and nothing will break. 🔄

GAME has many key advantages for genomics model benchmarking:
1) Via the API, all models are inherently compatible with all benchmarks
2) GAME will work across platforms – just get Apptainer installed, and away you go 💻↔️🖥️

Our Matcher containerizes an LLM, which is the basis for matching. It worked surprisingly well almost out of the box, and can match cell/tissue types, species, and molecule types. 🔮

The “Matcher” enables pairing up of the benchmark tasks and things the model can predict. Evaluator wants predictions in heart cells? The Matcher can tell you that your cardiomyocyte model is your best bet. The Matcher and Predictors also communicate via an API!

Benchmarks are containerized as “Evaluators”. We created several, including chromatin conformation, MPRA, and synthetic cis-regulatory variant effects. As new datasets come out, they can be added to GAME to see whether models can predict them.

Models are containerized in “Predictors”. We created DREAM-RNN (K562), Enformer, Orca, DeepBICCN2, and Borzoi Predictors. We anticipate as new models are created, model builders will encapsulate their models, referencing the examples we provide.

Diagram of the GAME modules and how they interact.

GAME uses Application Programming Interfaces, a concept introduced to the genomics model world with Kipoi, to enable uniform application of any benchmark to any model.

Unlike the protein folding code, gene regulation cannot have a single benchmark (e.g. CASP) because the gene regulatory code differs across species, cell types, and conditions. We need a variety of benchmarks to understand model strengths and weaknesses: enter, GAME.

Trying my best to channel first author @ishikaluthra.bsky.social for this, who is preoccupied this week… 🔮 You should follow her for updates!

SOOOO MANY GENOMICS MODELSSSS! 😱 Often unclear which is best since they benchmark differently! In this preprint, we introduce GAME, a new framework that utilizes APIs to enable sustainable, uniform model evaluation so we can see which is actually best for each task. doi.org/10.1101/2025...

n++;

Try reading that study. It appears to be very poorly done. Also just bizarre in many places. I wouldn't be surprised if it turns out to be total BS.

I think most are continuing to share it because it fits their preconceptions.

Lots of great things start at conferences in Barbados! What a story, Jeff! Had no idea. Sharing is much appreciated, and a great showcase of the value of biomedical research, especially needed right now.

A multi-kingdom genetic barcoding system for precise clone isolation - Nature Biotechnology A barcoded CRISPR base editing system isolates target clones from complex mammalian, yeast and bacterial populations.

CloneSelect published in Nature Biotechnology
@natbiotech.nature.com. This retrospective clone isolation method using CRISPR base editors is a powerful tool in broad biology. A history of Soh in the Yachie lab. www.nature.com/articles/s41...

🎉 This paper has been a long time and a labour of love (and hardship) for multiple group members, but, finally: we MPRA'ed 25k introgressed variants (Denisovan and Neanderthal) segregating at allele frequencies > 0.15 in humans today to evaluate their potential to regulate gene expression.

DNA-guided transcription factor interactions extend human gene regulatory code - Nature A large-scale analysis of DNA-bound transcription factors (TFs) shows how the presence of DNA markedly affects the landscape of TF interactions, and identifies composite motifs that are recognized by ...

A tour de force study from Taipale&Yin labs. It expands the vocabulary of the Regulatory Code by adding 1131 TF:TF composite motifs that are different from the individual TF motifs. The new composite motifs are enriched in cell-type specific elements and active in vivo
www.nature.com/articles/s41...

CAGT Poster featuring Keynote, Dr. Calin Plesa, of DropSynth fame!

⏰Abstract deadline for the Cascadia Advanced Genomic Technologies Meeting is April 30!⚠️ We're hoping this will be the first of many, catalyzing collaboration and leveraging our regional strength in advanced genomics technologies! See you there!
de-boer-lab.github.io/CAGT_meeting/

Join us for the 2025 SynBio7.0 conference in Toronto, June 1-3!

SynBio is Canada's largest academic synthetic biology conference.

More details on our website! synbio7.vercel.app

Our latest work now online in Cell:

Rewriting regulatory DNA to dissect and reprogram gene expression

Our new method (Variant-EFFECTS) uses high-throughput prime editing + flow sorting + sequencing to precisely measure effects of noncoding variants on gene expression

Thread 👇

Poster for CAGT, featuring Keynote from Dr. Calin Plesa, U Oregon

Reminder: abstract deadline for the Cascadia Advanced Genomic Technologies Meeting is April 30! We're hoping this will be the first of many, catalyzing collaboration and leveraging our regional strength in advanced genomics technologies! Please spread the word!
de-boer-lab.github.io/CAGT_meeting/

Latest preprint from the lab describing a super fast way to clone arrays of CRISPR guide RNAs. If you want to target multiple sites simultaneously, and want fast iteration on building and testing arrays of guides, look no further!
Bonus: it's named after a Pokemon🔥🦄

From April 14-18, select UBC graduate programs at UBC Vancouver will re-open their applications for US citizens to be considered for September 2025 or January 2026 entry - they are ready to provide quick admissions decisions for these applicants. If you are a prospective graduate student from the United States considering Canada for graduate school – now is the time.

UBC is hosting US Applicant Week from April 14-18, which allows US students to apply during an extended application period for graduate studies. Sixty graduate programs will be reopening applications for US applicants for September 2025 and January 2026 start dates - www.grad.ubc.ca/us-applicant...

Openness guides discovery
@naturebiotech.bsky.social

"To paraphrase Dwight D. Eisenhower’s aphorism: in research, planning is indispensable, but detailed plans are useless".

Spot on by @itaiyanai.bsky.social and MJ Lercher

www.nature.com/articles/s41...

CREsted: modeling genomic and synthetic cell type-specific enhancers across tissues and species Sequence-based deep learning models have become the state of the art for the analysis of the genomic regulatory code. Particularly for transcriptional enhancers, deep learning models excel at decipher...

Very proud of two new preprints from the lab:
1) CREsted: to train sequence-to-function deep learning models on scATAC-seq atlases, and use them to decipher enhancer logic and design synthetic enhancers. This has been a wonderful lab-wide collaborative effort. www.biorxiv.org/content/10.1...