Danny Jesus Diaz, PhD
I am a computational protein engineer.
My research consists of developing sequence- and structure-based machine learning frameworks
for identifying stabilizing and functional mutations in proteins. I collaborate extensively with
experimental protein engineers to accelerate the developability and functionality of proteins for
biotechnology applications.
I received my PhD in Chemistry under Dr Andrew Ellington,
and Dr Eric Anslyn at the
University of Texas at Austin.
During my PhD, I was the primary developer of MutCompute:
a machine learning as a service tool for structure-based ML-guided protein engineering.
Currently, I lead the Deep Proteins Groups at
the Institute for Foundations of Machine Learning (IFML).
I co-founded Intelligent Proteins, LLC where we use machine learning-guided protein engineering
to develop protein-based biotechnologies for nutraceutical, therapeutic, and biomanufacturing
applications.
Email /
CV /
Bio /
LinkedIn /
Google Scholar
/
Twitter /
Github
|
|
Research
I'm interested in protein engineering, machine learning, computer vision, biocatalysis,
cancer metabolism, rare metabolic diseases, automation, and startup/entrepreneurship.
My research consist of developing machine learning frameworks that leverage sequence, structure, and/or functional data for enzyme discovery and protein engineering applications.
Representative papers are highlighted.
|
Machine Learning Papers
|
A Systematic Evaluation of The Language-of-Viral-Escape Model Using Multiple Machine Learning Frameworks
Brent Allman,
Luiz Vieira,
Daniel J Diaz,
Claus O Wilke,
BioRxiv, 2024
It is critical to rapidly identify mutations with the potential for immune escape or
increased disease burden (variants of concern). A recent study
proposed that viral variants-of-concern can be identified using two quantities extracted from protein language models: grammaticality and semantic change.
Grammaticality is intended to be a measure of whether a viral protein variant is viable, and semantic change is
intended to be a measure of the variants potential for immune escape. Here, we systematically test this hypothesis, taking advantage of
several high-throughput datasets that have become available since the original study, and also evaluating additional machine learning models for
calculating the grammaticality and semantic metrics. We find that grammaticality correlates with protein viability, though the
more traditional metric, ÎÎG, appears to be more effective. By contrast, we do not find compelling evidence that the semantic change metric
can effectively identifying immune escape mutations.
|
Evolution-Inspired Loss Functions for Protein Representation Learning
Chengyue_Gong,
Adam Klivans,
James Madigan Loy,
Tianlong Chen,
Qiang Liu,
Daniel J Diaz
International Conference on Machine Learning, 2024
Current, protein representation learning methods primarily rely on BERT- or GPT-style self-supervised learning
and use wildtype accuracy as the primary training/validation metric. Wildtype accuracy, however, does not align
with the primary goal of protein engineering: suggest beneficial mutations rather than to identify
what already appears in nature. To address this gap between the pre-training objectives and protein engineering downstream tasks,
we present Evolutionary Ranking (EvoRank): a training objective that incorporates evolutionary information derived from
multiple sequence alignments (MSAs) to learn protein representations specific for protein engineering applications.
Across a variety of phenotypes and datasets, we demonstrate that an EvoRank pre-trained graph-transformer (MutRank) results
in significant zero-shot performance improvements that are competitive with ML frameworks
fine-tuned on experimental data. This is particularly important in protein engineering, where it is expensive to obtain data for fine-tuning.
|
Stability Oracle: A Structure-Based Graph-Transformer for Identifying Stabilizing Mutations
Daniel J Diaz,
Chengyue Gong,
Jeffrey Ouyang-Zhang,
James M Loy,
Jordan Wells,
David Yang,
Andrew D Ellington,
Alexandros G Dimakis,
Adam R Klivans
Nature Communications, 2024
Stability Oracle is a structure-based graph-transformer framework that is first pre-trained with BERT-style self-supervision on
the MutComputeX dataset and then fine-tuned on a curated subset of
the megascale cDNA-display proteolysis dataset.
Here, we introduce several innovations to overcome well-known challenges in data scarcity
and bias, generalization, and computation time, such as: Thermodynamic Permutations for data augmentation,
structural amino acid embeddings to model a mutation with a single structure,
a protein structure-specific attention-bias mechanism that makes graph transformers a viable alternative
to graph neural networks.
|
Binding Oracle: Fine-Tuning From Stability to Binding Free Energy
Chengyue Gong,
Adam R Klivans,
Jordan Wells,
James Loy,
Qiang Liu,
Alexandros G Dimakis,
Daniel J Diaz,
NeurIPS GenBio Workshop Spotlight, 2023
Fine-tuning machine learning frameworks to a small experimental dataset is prone to overfitting.
Here, we present Binding Oracle: a Graph-Transformer framework that fine-tunes Stability Oracle
to ââG of binding for protein-protein interfaces (PPI) via a technique we call Selective LoRA.
Selective LoRA, uses the gradient norms of each layer to select the subset most sensitive to the
fine-tuning dataset--here it was a curated subset of Skempi2.0 (B1816)--and then finetunes the
selected layers with LoRA. By applying Selective LoRA to Stability Oracle, we are able to achieve
SOTA on the S487 PPI test set and generalization between different types of PPI interfaces.
|
Predicting a Proteinâs Stability under a Million Mutations
Jeffrey Ouyang-Zhang,
Daniel J Diaz,
Adam R Klivans
Philipp Krahenbuhl
NeurIPS, 2023
The mutate everything framework allows the fine-tuning of sequence-based (ESM2) and MSA-based
(AlphaFold2) protein foundation models on phenotype data with parallel decoding. Here, we
demonstrate how their
representations can be fine-tuned on the
cDNA-display proteolysis
dataset with the mutate everything framework to predict the thermodynamic impact of
single point mutations (ââG). The AlphaFold2 fine-tuned model, StabilityFold, is able to achieve
similar
results to Stability Oracle on a variety test sets. More importantly, The mutate everything
framework allows for parallel decoding of single and higher-order amino acid substitutions into ââG
predictions.
This capability not only enables rapid DMS inferencing of proteins but makes double
mutant DMS inferencing computationally tractable.
|
Hotprotein: A novel framework for protein thermostability prediction and editing
Tianlong Chen,
Chengyue Gong,
Daniel J Diaz,
Xuxi Chen,
Jordan Tyler Wells,
Zhangyang Wang,
Andrew Ellington,
Alexandros G Dimakis,
Adam Klivans
ICLR, 2023
We curated an organism-based temperature dataset (HotProteins) for distinguishing proteins
with varying thermostability (cryophiles, psychrophiles, mesophiles, thermophiles, and
hyperthermophiles).
We proposed structure-aware pretraining (SAP) and factorized sparse tuning (FST) to
fine-tune the sequence-based transformer, ESM-1b, representations to generate a classifier and
regressor to predict a protein's organism class or growth temperature.
|
Two sequence- and two structure-based ML models have learned different aspects of protein
biochemistry
Anastasiya V Kulikova,
Daniel J Diaz,
Tianlong Chen,
Jeffrey Cole,
Andrew D Ellington,
Claus O Wilke
Scientific Reports, 2023
We compare and contrast self-supervised sequence-based transformers and structure-based 3DCNNs
models.
We find that there is a variance-bias tradeoff between the two protein modalities.
Convolutions provide an inductive bias for protein structures where the
more powerful sequence-based transformers demonstrate increase variance.
|
Learning the Local Landscape of Protein Structures with Convolutional Neural Networks
Anastasiya V Kulikova,
Daniel J Diaz,
James M Loy,
Andrew D Ellington,
Claus O Wilke
Journal of Biological Physics, 2021
We compare how self-supervised 3DCNNs learn the local mutational landscape of proteins
against evolution via Multiple Sequence Alignments. We find that structure-based 3DCNNs
amino acid likelihoods have weak correlation with MSAs and their wildtype confidence
is dependent on the structural position of the residue.
Where core residues being more confidently predicted.
|
Protein Papers
|
Engineering a photoenzyme to use red light
Jose M. Carceller,
Bhumika Jayee,
Claire G. Page,
Daniel G. Oblinsky,
Gustavo MondragĂłn-SolĂłrzano,
Nithin Chintala,
Jingzhe Cao,
Zayed Alassad,
Zheyu Zhang,
Nathaniel White,
Daniel J. Diaz,
Andrew D. Ellington,
Gregory D. Scholes,
Sijia S. Dong,
Todd K. Hyster
Cell Chem, 2024
Previously, we engineered a ene-reductase (ERED) photoenzyme capable of
asymmetric synthesis of α-chloroamides under blue light using directed evolution
and machine learning-guided protein engineering with MutComputeX.
Here, we conduct allosteric tuning of the electronic structure of the cofactor-substrate complex using
directed evolution and MutComputeX to dramatically increase catalysis under red light (99% yields).
Computational studies show a different electron transition for cyan and red light and
how mutations at the protein surface allosterically tune the active site complex.
|
Biosensor and machine learning-aided engineering of an amaryllidaceae enzyme
Simon d'Oelsnitz,
Daniel J Diaz,
Daniel J Acosta,
Tyler L Dangerfield,
Mason W Schechter,
Matthew B Minus,
James R Howard,
Hannah Do,
James Loy,
Hal Alper,
Andrew D Ellington
Nature Communications, 2024
We engineered a transcription factor and a methyl transferase to improve the
regioselectivity and titer yield for production of 4O-methyl-norbelladine.
To engineer the 4O-methyltransferase enzyme, we developed MutComputeX:
a self-supervised 3DResNet trained to generalize to protein-ligand, -nucleotide, and -protein
interfaces.
MutComputeX was used to design mutations on a computational ternary structure of
the AlphaFolded methyl-transferase
with SAH and norbelladine docked
with Gnina.
This is the first time three machine learning models (AlphaFold, Gnina, MutComputeX) were
synergized to engineer the surface and active site of an enzyme and combined with an
engineered transcription factor for high-throughput screening.
|
Asymmetric Synthesis of α-Chloroamides via Photoenzymatic Hydroalkylation of Olefins
Yi Liu,
Sophie G Bender,
Damien Sorigue,
Daniel J Diaz,
Andrew D Ellington,
Greg Mann,
Simon Allmendinger,
Todd K Hyster
Journal of the American Chemical Society, 2024
We engineered a Ene-reductase photoenzyme capable of asymmetric synthesis of α-chloroamides under blue light.
MutComputeX was used to identify several mutation that improved the activity and stereoselectivity
observed under blue light.
|
Machine learning-aided engineering of hydrolases for PET depolymerization
Hongyuan Lu,
Daniel J Diaz,
Natalie J Czarnecki,
Congzhi Zhu,
Wantae Kim,
Raghav Shroff,
Daniel J Acosta,
Bradley R Alexander,
Hannah O Cole,
Yan Zhang,
Nathaniel A Lynd,
Andrew D Ellington,
Hal S Alper
Nature, 2022
We utilized MutCompute to guide the engineering of a mesophilic and thermophilic PET hydrolases.
We examined the ability of the mesophilic PET hydrolase (FAST-PETase) to depolymerize post-consumer
PET waste.
FAST-PETase was capable of depolymerizing ~50 post-consumer PET waste within 2-4 days. Furthermore,
the ML designs
increased the depolymerization capacity of the thermophilic PET hydrolase (ICCM) by 100%.
Here is a time-lapse video of the depolymerization of a
full PET container from Walmart.
|
Improved Bst DNA Polymerase Variants Derived via a Machine Learning Approach
Inyup Paik,
Phuoc HT Ngo,
Raghav Shroff,
Daniel J Diaz,
Andre C Maranhao,
David JF Walker,
Sanchita Bhadra,
Andrew D Ellington
Biochemistry, 2021
Bst Polymerase was stabilized via ML-guided protein engineering in order to shorten the diagnostic
time of LAMP-OSD assays during the height of the COVID19 pandemic. LAMP-OSD is an isothermal DNA
amplification technique that enables field diagnostic of COVID19 in poor resource settings.
|
GroovDB: A database of ligand-inducible transcription factors
Simon dâOelsnitz,
Joshua D Love,
Daniel J Diaz,
Andrew D Ellington
ACS SynBio, 2022
A database for ligand-induced transcription factor.
These transcription factor serve as starting points for high-throughput screening genetic
biosensors.
|
Discovery of novel gain-of-function mutations guided by structure-based deep learning
Raghav Shroff,
Austin W Cole,
Daniel J Diaz,
Barrett R Morrow,
Isaac Donnell,
Ankur Annapareddy,
Jimmy Gollihar,
Andrew D Ellington,
Ross Thyer
ACS SynBio, 2020
The development and initial experimental validation of the MutCompute framework: a self-supervised
3DCNN trained on the local chemistry surrounding an amino acid. The model was experimentally
characterized for its ability to identify residues where the wildtype amino acid is incongruent for
its surrounding chemical environment (protein only) and primed for gain-of-function.
Here, BFP, phosphomannose isomerase, and beta-lactamase were engineered via machine learning.
|
Reviews
|
Using machine learning to predict the effects and consequences of mutations in proteins
Daniel J Diaz,
Anastasiya V Kulikova,
Andrew D Ellington,
Claus O Wilke
Current Opinion in Structural Biology, 2023
A review on the state-of-the-art machine learning frameworks (as of July 2022) for characterizing
the
functional and stability effects of point mutations on proteins.
|
Pushing Differential Sensing Further: The Next Steps in Design and Analysis of BioâInspired
CrossâReactive Arrays
Hazel A Fargher,
Simon d'Oelsnitz,
Daniel J Diaz,
Eric V Anslyn
Analysis & Sensing, 2023
A perspective on the future technology developments of differential sensing.
|
Patents
|
Mutations for improving activity and thermostability of petase enzymes
Hongyuan Lu,
Daniel J Diaz,
Hannah Cole,
Raghav Shroff,
Andrew D Ellington,
Hal Alper
WO Patent WO2022076380A2
ML-engineered PETase variants, inluding the FAST-PETase varaint. Discovery of the N233K ML-design
that replaces Ca2+ dependence for stability and activity with a lysine cation.
|
Leaf-branch compost cutinase mutants
Hongyuan Lu,
Daniel J Diaz,
Andrew D Ellington,
Hal Alper
WO Patent WO2023154690A2
ML-engineered Cutinase variant, which makes use of the N233K ML-design
that replaces Ca2+ dependence for stability and activity with a lysine cation.
|
Engineered human serine dehydratase enzymes and methods for treating cancer
Everett Stone,
Ebru Cayir,
Daniel J Diaz,
Raghav Shroff
WO Patent WO2024006973A1
ML-engineered variants of the human serine dehydratase enzyme for the treatment of luminal breast cancer.
|
Recombinant proteins with increased solubility and stability
Andrew Ellington,
Inyup Paik,
Andre Maranhao,
Sanchita Bhadra,
David Walker,
Daniel J Diaz,
Ngo Phuoc
US Patent US20240011000A1
ML-engineered Bst DNA Polymerase variants with increased solubility and stability for LAMP-OSD assays for isothermal COVID19 diagnostics.
|
Methods and compositions related to modified methyltransferases and engineered biosensors
Andrew Ellington,
Daniel J Diaz,
Simon d'Oelsnitz
US Patent App. 63/493,065
ML-engineered Norbelladine 4O-MethylTransferase variants for biomanufacturing of Galantamine intermediate 4O-methyl-norbelladine.
First, ML-engineered enzyme where the ML-designs were conditioned on the active site ligand and cofactor.
ML-designs were generated using a computational ternary structure: protein (AlphaFold2), SAH cofactor (GNINA-docking),
and norbelladine substrate (GNINA-docking).
|
|