Peter Koo
Assistant Professor
Cancer Center Member
Ph.D., Yale University, 2015
koo@cshl.edu | 516-367-5520
Deep learning has the potential to make a significant impact in basic biology and cancer, but a major challenge is understanding the reasons behind their predictions. My research develops methods to interpret this powerful class of black box models, with a goal of elucidating data-driven insights into the underlying mechanisms of sequence-function relationships.
Deep learning is being applied rapidly in many areas of genomics, demonstrating improved performance over previous methods on benchmark datasets. Despite the promise of deep learning, it remains unclear whether improved predictions will translate to new biological discoveries because of their low interpretability, which has earned them a reputation as a black box. Understanding the reasons underlying a deep learning model’s prediction may reveal new biological insights not captured by previous methods. Our group develops methods to interpret high-performing deep learning models to distill knowledge that they learn from big, noisy, biological sequence data. Our goal is to elucidate biological mechanisms that underlie sequence-function relationships for gene regulation and protein (dys)function. Recently, we have teamed up with other members of the CSHL Cancer Center to investigate the sequence basis of epigenomic differences across healthy and cancer cells.
SQUID pries open AI black box
June 21, 2024
CSHL’s Koo and Kinney labs have built a tool to suss out how AI analyzes the genome. What sets it apart? Decades of quantitative genetics knowledge.
The CSHL School of Biological Sciences’ class of 2024
May 5, 2024
The School of Biological Sciences awarded Ph.D. degrees to 11 students this year. Here are some stories and reflections from their time at CSHL.
Can AI uncover breast cancer risk factors?
February 26, 2024
This question lies at the heart of a new interdisciplinary collaboration between CSHL’s Camila dos Santos and Peter Koo.
The digital dark matter clouding AI
June 5, 2023
Scientists have unknowingly encountered mysterious noise while using AI to decipher our genetic code. CSHL has found a way to cut through the fog.
AI training: A backward cat pic is still a cat pic
May 4, 2023
This basic rule of thumb is helping CSHL’s quantitative biologists train AI to get a better read of the human genome.
Can you outsmart this AI quiz?
February 6, 2023
Think you’re plugged into the latest artificial intelligence advancements? Test your tech knowledge with this quiz on AI and computational biology.
Finding the right AI for you
December 5, 2022
AI’s popularity has reached a point where there are too many options. How do you know which AI is right for you? CSHL scientists have a solution.
CSHL high schoolers finish top 10 in 2022 DREAM Challenge
October 7, 2022
The high school team competed against universities and private labs to build a computer program for predicting gene expression in yeast.
Regeneron competition honors CSHL high school researchers
March 22, 2022
Three high school student researchers at CSHL were among Regeneron Science Talent Search’s top 300 scholars. One made it to the final competition.
Making AI algorithms show their work
May 13, 2021
AI machines are often better than humans at discerning patterns. CSHL researchers developed a way to find out why.
All Publications
Interpretably deep learning amyloid nucleation by massive experimental quantification of random sequences
17 Jul 2024 | Cold Spring Harbor Laboratory
Thompson, Mike, Martin, Mariano, Olmo, Trinidad, Rajesh, Chandana, Koo, Peter, Bolognesi, Benedetta, Lehner, Ben
Interpreting cis-regulatory mechanisms from genomic deep neural networks using surrogate models
Jun 2024 | Nature Machine Intelligence | 6(6):701-713
Seitz, E, McCandlish, D, Kinney, J, Koo, P
Designing DNA With Tunable Regulatory Activity Using Discrete Diffusion
24 May 2024 | bioRxiv
Sarkar, Anirban, Tang, Ziqi, Zhao, Chris, Koo, Peter
Evaluating the representational power of pre-trained DNA language models for regulatory genomics
4 Mar 2024 | bioRxiv
Tang, Ziqi, Koo, Peter
EvoAug-TF: Extending evolution-inspired data augmentations for genomic deep learning to TensorFlow
16 Feb 2024 | Bioinformatics
Yu, Yiyang, Muthukumar, Shivani, Koo, Peter, Martelli, Pier
EvoAug-TF: Extending evolution-inspired data augmentations for genomic deep learning to TensorFlow
18 Jan 2024 | bioRxiv
Yu, Yiyang, Muthukumar, Shivani, Koo, Peter
Current approaches to genomic deep learning struggle to fully capture human genetic variation
Dec 2023 | Nature Genetics | 55(12):2021-2022
Tang, Ziqi, Toneyan, Shushan, Koo, Peter
Interpreting cis -regulatory mechanisms from genomic deep neural networks using surrogate models
16 Nov 2023 | bioRxiv
Seitz, Evan, McCandlish, David, Kinney, Justin, Koo, Peter
ChampKit: A framework for rapid evaluation of deep neural networks for patch-based histopathology classification
Sep 2023 | Computer Methods and Programs in Biomedicine | 239:107631
Kaczmarzyk, Jakub, Gupta, Rajarsi, Kurc, Tahsin, Abousamra, Shahira, Saltz, Joel, Koo, Peter
Interpreting Cis -Regulatory Interactions from Large-Scale Deep Neural Networks for Genomics
3 Jul 2023 | bioRxiv
Toneyan, Shushan, Koo, Peter