arXiv subjects

Vladimir Gligorijević

Publications and source records attributed to Vladimir Gligorijević.

5 recordsLinked to original sources

arXiv2025

Generalists vs. Specialists: Evaluating LLMs on Highly-Constrained Biophysical Sequence Optimization Tasks

Although large language models (LLMs) have shown promise in biomolecule optimization problems, they incur heavy computational costs and struggle to satisfy precise constraints. On the other hand, specialized solvers like LaMBO-2 offer efficiency and fine-grained control but require more domain expertise. Comparing these approaches is challenging due to expensive laboratory validation and inadequate synthetic benchmarks. We address this by introducing Ehrlich functions, a synthetic test suite that captures the geometric structure of biophysical sequence optimization problems. With prompting alone, off-the-shelf LLMs struggle to optimize Ehrlich functions. In response, we propose LLOME (Language Model Optimization with Margin Expectation), a bilevel optimization routine for online black-box optimization. When combined with a novel preference learning loss, we find LLOME can not only learn to solve some Ehrlich functions, but can even perform as well as or better than LaMBO-2 on moderately difficult Ehrlich variants. However, LLMs also exhibit some likelihood-reward miscalibration and struggle without explicit rewards. Our results indicate LLMs can occasionally provide significant benefits, but specialized solvers are still competitive and incur less overhead.

cs.LG↗

arXiv2024

Implicitly Guided Design with PropEn: Match your Data to Follow the Gradient

Across scientific domains, generating new models or optimizing existing ones while meeting specific criteria is crucial. Traditional machine learning frameworks for guided design use a generative model and a surrogate model (discriminator), requiring large datasets. However, real-world scientific applications often have limited data and complex landscapes, making data-hungry models inefficient or impractical. We propose a new framework, PropEn, inspired by ``matching'', which enables implicit guidance without training a discriminator. By matching each sample with a similar one that has a better property value, we create a larger training dataset that inherently indicates the direction of improvement. Matching, combined with an encoder-decoder architecture, forms a domain-agnostic generative framework for property enhancement. We show that training with a matched dataset approximates the gradient of the property of interest while remaining within the data distribution, allowing efficient design optimization. Extensive evaluations in toy problems and scientific applications, such as therapeutic protein design and airfoil optimization, demonstrate PropEn's advantages over common baselines. Notably, the protein design results are validated with wet lab experiments, confirming the competitiveness and effectiveness of our approach.

cs.LG↗

arXiv2022

A Pareto-optimal compositional energy-based model for sampling and optimization of protein sequences

Deep generative models have emerged as a popular machine learning-based approach for inverse design problems in the life sciences. However, these problems often require sampling new designs that satisfy multiple properties of interest in addition to learning the data distribution. This multi-objective optimization becomes more challenging when properties are independent or orthogonal to each other. In this work, we propose a Pareto-compositional energy-based model (pcEBM), a framework that uses multiple gradient descent for sampling new designs that adhere to various constraints in optimizing distinct properties. We demonstrate its ability to learn non-convex Pareto fronts and generate sequences that simultaneously satisfy multiple desired properties across a series of real-world antibody design tasks.

cs.LG↗

arXiv2022

Multi-segment preserving sampling for deep manifold sampler

Deep generative modeling for biological sequences presents a unique challenge in reconciling the bias-variance trade-off between explicit biological insight and model flexibility. The deep manifold sampler was recently proposed as a means to iteratively sample variable-length protein sequences by exploiting the gradients from a function predictor. We introduce an alternative approach to this guided sampling procedure, multi-segment preserving sampling, that enables the direct inclusion of domain-specific knowledge by designating preserved and non-preserved segments along the input sequence, thereby restricting variation to only select regions. We present its effectiveness in the context of antibody design by training two models: a deep manifold sampler and a GPT-2 language model on nearly six million heavy chain sequences annotated with the IGHV1-18 gene. During sampling, we restrict variation to only the complementarity-determining region 3 (CDR3) of the input. We obtain log probability scores from a GPT-2 model for each sampled CDR3 and demonstrate that multi-segment preserving sampling generates reasonable designs while maintaining the desired, preserved regions.

cs.LG↗

arXiv2014

FUSE: Multiple Network Alignment via Data Fusion

Discovering patterns in networks of protein-protein interactions (PPIs) is a central problem in systems biology. Alignments between these networks aid functional understanding as they uncover important information, such as evolutionary conserved pathways, protein complexes and functional orthologs. The objective of a multiple network alignment is to create clusters of nodes that are evolutionarily conserved and functionally consistent across all networks. Unfortunately, the alignment methods proposed thus far do not fully meet this objective, as they are guided by pairwise scores that do not utilize the entire functional and topological information across all networks. To overcome this weakness, we propose FUSE, a multiple network aligner that utilizes all functional and topological information in all PPI networks. It works in two steps. First, it computes novel similarity scores of proteins across the PPI networks by fusing from all aligned networks both the protein wiring patterns and their sequence similarities. It does this by using Non-negative Matrix Tri-Factorization (NMTF). When we apply NMTF on the five largest and most complete PPI networks from BioGRID, we show that NMTF finds a larger number of protein pairs across the PPI networks that are functionally conserved than can be found by using protein sequence similarities alone. This demonstrates complementarity of protein sequence and their wiring patterns in the PPI networks. In the second step, FUSE uses a novel maximum weight k-partite matching approximation algorithm to find an alignment between multiple networks. We compare FUSE with the state of the art multiple network aligners and show that it produces the largest number of functionally consistent clusters that cover all aligned PPI networks. Also, FUSE is more computationally efficient than other multiple network aligners.

q-bio.MN↗