Skip to content

Navigation Menu

Sign in
Sign up
@EmilioVenegas
EmilioVenegas
Follow

Emilio Venegas EmilioVenegas

Computational biologist. Geometric deep learning for structure-based drug design; longitudinal clinical modeling. prev. Harvard/BWH.

Highlights

  • Pro

Block or report EmilioVenegas

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.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
EmilioVenegas /README.md

Emilio Venegas

Computational biologist working on geometric deep learning for structure-based drug design and longitudinal modeling of clinical cohorts.
B.Sc. Biosciences, Tecnológico de Monterrey (2026) · Guadalajara, Mexico

Computational Biology Portfolio


01 // Selected Work

SBDD 3D Diffusion Benchmarking

Published 3D generative models (DiffSBDD, TargetDiff, Pocket2Mol) delete catalytic metal ions during preprocessing, as identified by inspecting their source code. On a curated 133-target zinc benchmark they coordinate zinc validly in 20% of molecules against a 77% native-ligand ceiling, and fine-tuning on more metalloprotein data makes it worse, not better. Analysis plans pre-registered before the data; repair in progress.

02 · strider

Biophysics Thermodynamics PyPI: strider-dna

Nucleic acid thermodynamics and kinetics in pure Python: nearest-neighbor free energies, O(n3) folding (Zuker, McCaskill), strand-displacement rate constants, parallel-tempered sequence design.

03 · cadence

Neural ODE Longitudinal VAE Clinical Cohorts

Latent Neural ODE-VAE over six waves of the Mexican Health and Aging Study; 34 clinical deficits compressed to an 8D manifold with aging velocity as output. Cognitive-decline AUC 0.72 → 0.79, functional 0.67 → 0.75, patient-level 5-fold CV.

Foundation Models SE(3)-GNN Antibiotic Design

De novo antibiotic design against multidrug-resistant A. baumannii, P. aeruginosa, K. pneumoniae. Conditions DiffSBDD on ATOMICA pocket embeddings through an SE(3)-equivariant cross-graph convolution, with an evolutionary RL loop over ADMET endpoints.

CRNT Dynamical Systems SymPy

Chemical reaction network theory under mass-action kinetics: network invariants, symbolic ODEs and Jacobians via SymPy, stability certified from topology alone.


02 // Technical Toolbox

  • Machine Learning & AI: PyTorch · Graph Neural Networks · SE(3)-equivariant GNNs (e3nn, EGNN) · Diffusion models (DiffSBDD) · Neural ODEs · VAEs · Reinforcement Learning · Foundation Models (ATOMICA) · Multiple Instance Learning (MIL) · Vision Transformers · scikit-learn · TensorFlow
  • Drug Discovery & Cheminformatics: RDKit · Structure-based drug design · AutoDock Vina · Boltz-2 · ADMET-AI · CrossDocked curation · SMILES · Tanimoto similarity · REOS filters · Lipinski rules
  • LLM & RAG Systems: Retrieval-Augmented Generation (dense/lexical RRF, cross-encoders) · Vector databases (pgvector) · Agentic routing · LangChain · LangGraph · Hugging Face Transformers
  • Software & Systems: Python · C++ · R · SQL · JavaScript (React) · Flask · Docker · Git · Linux · GPU/HPC clusters · Embedded Systems (Raspberry Pi/ESP32)
  • Bioinformatics & Wet Lab: Genomic Data Analysis · ImageJ/Fiji · Chimera · Cell culture · 3D spheroids · PCR/qPCR · Flow cytometry · Western blotting · Organic synthesis

03 // Elsewhere & Honors

Pinned Loading

  1. metal-aware-sbdd metal-aware-sbdd Public

    Benchmarking and repairing metal coordination failures in pocket-conditioned 3D generative models for structure-based drug design (SBDD).

    Python 1

  2. strider strider Public

    A Python library for nucleic acid thermodynamics, kinetics, and DNA/RNA circuit design with end-to-end simulation pipelines.

    Python 4 1

  3. mantis-delta mantis-delta Public

    A Python library for Chemical Reaction Network Theory (CRNT) bridging network topology with symbolic mass-action kinetics and algebraic stability certification.

    Python 1

  4. cadence cadence Public

    Latent Neural ODE-VAE for continuous aging dynamics: 34 clinical deficits from six waves of the Mexican Health and Aging Study compressed into an 8-dimensional manifold whose native output is aging...

    Python

  5. ATOMICA-Diffusion-Antibiotic-design ATOMICA-Diffusion-Antibiotic-design Public

    ATOMICA-Diffusion-RL is a unified pipeline for de novo antibiotic design. It conditions an SE(3)-equivariant diffusion model (DiffSBDD) using ATOMICA’s universal interaction interface embeddings. T...

    Python 1 1

  6. fda-rag-system fda-rag-system Public

    Python

AltStyle によって変換されたページ (->オリジナル) /