Improving functional protein generation via foundation model-derived latent space likelihood optimization DOI Creative Commons
Changge Guan, Fangping Wan, Marcelo D. T. Torres

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Янв. 8, 2025

A variety of deep generative models have been adopted to perform de novo functional protein generation. Compared 3D design, sequence-based generation methods, which aim generate amino acid sequences with desired functions, remain a major approach for due the abundance and quality sequence data, as well relatively low modeling complexity training. Although these are typically trained match from training exact matching every is not always essential. Certain changes (e.g., mismatches, insertions, deletions) may necessarily lead changes. This suggests that maximizing data likelihood beyond space could yield better models. Pre-trained large language (PLMs) like ESM2 can encode into latent space, potentially serving validators. We propose by simultaneously optimizing in both derived PLM. scheme also be viewed knowledge distillation dynamically re-weights samples during applied our method train GPT- (i.e., autoregressive transformers) antimicrobial peptide (AMP) malate dehydrogenase (MDH) tasks. Computational experiments confirmed outperformed various adversarial net, variational autoencoder, GPT model without proposed strategy) on tasks, demonstrating effectiveness multi-likelihood optimization strategy.

Язык: Английский

Overview of AlphaFold2 and breakthroughs in overcoming its limitations DOI
Lei Wang,

Zehua Wen,

Shiwei Liu

и другие.

Computers in Biology and Medicine, Год журнала: 2024, Номер 176, С. 108620 - 108620

Опубликована: Май 15, 2024

Язык: Английский

Процитировано

13
AlphaFold two years on: Validation and impact DOI Creative Commons
Oleg Kovalevskiy, Juan Mateos-García, Kathryn Tunyasuvunakool

и другие.

Proceedings of the National Academy of Sciences, Год журнала: 2024, Номер 121(34)

Опубликована: Авг. 12, 2024

Two years on from the initial release of AlphaFold, we have seen its widespread adoption as a structure prediction tool. Here, discuss some latest work based with particular focus use within structural biology community. This encompasses cases like speeding up determination itself, enabling new computational studies, and building tools workflows. We also look at ongoing validation predictions continue to be compared against large numbers experimental structures further delineate model’s capabilities limitations.

Язык: Английский

Процитировано

12
Peptides of a Feather: How Computation Is Taking Peptide Therapeutics under Its Wing DOI Open Access
Tom Kazmirchuk,

Calvin Bradbury-Jost,

Taylor Ann Withey

и другие.

Genes, Год журнала: 2023, Номер 14(6), С. 1194 - 1194

Опубликована: Май 29, 2023

Leveraging computation in the development of peptide therapeutics has garnered increasing recognition as a valuable tool to generate novel for disease-related targets. To this end, transformed field design through identifying that exhibit enhanced pharmacokinetic properties and reduced toxicity. The process

Язык: Английский

Процитировано

17
Challenges in computational discovery of bioactive peptides in ’omics data DOI Creative Commons
Luís Pedro Coelho, Célio Dias Santos Júnior, César de la Fuente‐Núñez

и другие.

PROTEOMICS, Год журнала: 2024, Номер 24(12-13)

Опубликована: Март 8, 2024

Peptides have a plethora of activities in biological systems that can potentially be exploited biotechnologically. Several peptides are used clinically, as well industry and agriculture. The increase available 'omics data has recently provided large opportunity for mining novel enzymes, biosynthetic gene clusters, molecules. While these primarily consist DNA sequences, other types provide important complementary information. Due to their size, the approaches proven successful at discovering proteins canonical size cannot naïvely applied discovery peptides. encoded directly genome short open reading frames (smORFs), or they derived from larger by proteolysis. Both peptide classes pose challenges simple methods prediction result numbers false positives. Similarly, functional annotation proteins, traditionally based on sequence similarity infer orthology then transferring functions between characterized uncharacterized ones, sequences. use techniques is much more limited alternative machine learning instead. Here, we review limitations traditional been developed bioactive with focus prokaryotic genomes metagenomes.

Язык: Английский

Процитировано

8
HighFold: accurately predicting structures of cyclic peptides and complexes with head-to-tail and disulfide bridge constraints DOI Creative Commons
Chenhao Zhang, Chengyun Zhang,

Tianfeng Shang

и другие.

Briefings in Bioinformatics, Год журнала: 2024, Номер 25(3)

Опубликована: Март 27, 2024

Abstract In recent years, cyclic peptides have emerged as a promising therapeutic modality due to their diverse biological activities. Understanding the structures of these and complexes is crucial for unlocking invaluable insights about protein target–cyclic peptide interaction, which can facilitate development novel-related drugs. However, conducting experimental observations time-consuming expensive. Computer-aided drug design methods are not practical enough in real-world applications. To tackles this challenge, we introduce HighFold, an AlphaFold-derived model study. By integrating specific details head-to-tail circle disulfide bridge structures, HighFold accurately predict complexes. Our demonstrates superior predictive performance compared other existing approaches, representing significant advancement structure–activity research. The openly accessible at https://github.com/hongliangduan/HighFold.

Язык: Английский

Процитировано

8
The limits of prediction: Why intrinsically disordered regions challenge our understanding of antimicrobial peptides DOI Creative Commons
Roberto Bello‐Madruga, Marc Torrent

Computational and Structural Biotechnology Journal, Год журнала: 2024, Номер 23, С. 972 - 981

Опубликована: Фев. 12, 2024

Antimicrobial peptides (AMPs) are molecules found in most organisms, playing a vital role innate immune defense against pathogens. Their mechanism of action involves the disruption bacterial cell membranes, causing leakage cellular contents and ultimately leading to death. While AMPs typically lack defined structure solution, they often assume conformation when interacting with membranes. Given this structural flexibility, we investigated whether intrinsically disordered regions (IDRs) AMP-like properties could exhibit antimicrobial activity. We tested 14 from different IDRs predicted have activity that nearly all them did not display anticipated effects. These failed adopt secondary had compromised membrane interactions, resulting hypothesize evolutionary constraints may prevent folding, even membrane-like environments, limiting their potential. Moreover, our research reveals current predictors fail accurately capture features dealing unstructured sequences. Hence, results presented here far-reaching implications for designing improving strategies therapies infectious diseases.

Язык: Английский

Процитировано

7
Identification of flavor peptides based on virtual screening and molecular docking from Hypsizygus marmoreuss DOI
Wenting Wang, Hongbo Li, Zhenbin Liu

и другие.

Food Chemistry, Год журнала: 2024, Номер 448, С. 139071 - 139071

Опубликована: Март 19, 2024

Язык: Английский

Процитировано

7
Highly conserved brain vascular receptor ALPL mediates transport of engineered viral vectors across the blood-brain barrier DOI Creative Commons

Tyler C. Moyer,

Brett Hoffman, Weitong Chen

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Март 14, 2024

Abstract Delivery of systemically administered therapeutics to the central nervous system (CNS) is restricted by blood-brain barrier (BBB). Bioengineered Adeno-Associated Virus (AAV) capsids have been shown penetrate BBB with great efficacy in mouse and non-human primate models, but their translational potential often limited species selectivity undefined mechanisms action. Here, we apply our RNA-guided TRACER AAV capsid evolution platform generate VCAP-102, an AAV9 variant markedly increased brain tropism following intravenous delivery both rodents primates. VCAP-102 demonstrates a similar CNS cynomolgus macaque, african green monkey, marmoset mouse, showing 20- 400-fold transgene expression across multiple regions relative AAV9. We demonstrate that enhanced results from direct interaction alkaline phosphatase (ALPL), highly conserved membrane-associated protein expressed on vasculature. interacts human, murine ALPL isoforms, ectopic sufficient initiate receptor-mediated transcytosis vitro transwell model. Our work identifies as cross-species gene vector strong for clinical translation establishes shuttle capable efficient transport maximize biotherapeutics.

Язык: Английский

Процитировано

6
Design of Cyclic Peptides Targeting Protein–Protein Interactions Using AlphaFold DOI Open Access
Takatsugu Kosugi, Masahito Ohue

International Journal of Molecular Sciences, Год журнала: 2023, Номер 24(17), С. 13257 - 13257

Опубликована: Авг. 26, 2023

More than 930,000 protein-protein interactions (PPIs) have been identified in recent years, but their physicochemical properties differ from conventional drug targets, complicating the use of small molecules as modalities. Cyclic peptides are a promising modality for targeting PPIs, it is difficult to predict structure target protein-cyclic peptide complex or design cyclic sequence that binds protein using computational methods. Recently, AlphaFold with offset has enabled predicting peptides, thereby enabling de novo designs. We developed enable structural prediction proteins and complexes found AlphaFold2 can structures high accuracy. also applied binder hallucination protocol AfDesign, method AlphaFold, we could predicted local-distance difference test lower separated binding energy per unit interface area native MDM2/p53 structure. Furthermore, was 12 other protein-peptide one complex. Our approach shows possible putative sequences PPI.

Язык: Английский

Процитировано

15
Applicability of AlphaFold2 in the modelling of coiled-coil domains DOI Creative Commons
Rafał Madaj, Mikel Martínez-Goikoetxea, Kamil Kamiński

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Март 10, 2024

Abstract Coiled coils are a common protein structural motif involved in cellular functions ranging from mediating protein-protein interactions to facilitating processes such as signal transduction or regulation of gene expression. They formed by two more alpha helices that wind around central axis form buried hydrophobic core. Various forms coiled-coil bundles have been reported, each characterized the number, orientation, and degree winding constituent helices. This variability is underpinned short sequence repeats coiled whose properties determine both their overall topology local geometry The strikingly repetitive has enabled development accurate sequence-based prediction methods; however, modeling domains remains challenging task. In this work, we evaluated accuracy AlphaFold2 domains, predicting global topological properties. Furthermore, show oligomeric state can be achieved using internal representations AlphaFold2, with performance better than any previous state-of-the-art method (code available at https://github.com/labstructbioinf/dc2_oligo ).

Язык: Английский

Процитировано

5