Nat Prod Rep. 2023 Feb 28. doi: 10.1039/d2np00077f. Online ahead of print.

ABSTRACT

Glycosylation is a successful strategy to alter the pharmacological properties of small molecules, and it has emerged as a unique approach to expand the chemical space of natural products that can be explored in drug discovery. Traditionally, most glycosylation events have been carried out chemically, often requiring many protection and deprotection steps to achieve a target molecule. Enzymatic glycosylation by glycosyltransferases could provide an alternative strategy for producing new glycosides. In particular, the glycosyltransferase family has greatly expanded in plants, representing a rich enzymatic resource to mine and expand the diversity of glycosides with novel bioactive properties. This article highlights previous and prospective uses for plant glycosyltransferases in generating bioactive glycosides and altering their pharmacological properties.

PMID:36853278 | DOI:10.1039/d2np00077f

Elife. 2023 Feb 28;12:e83398. doi: 10.7554/eLife.83398. Online ahead of print.

ABSTRACT

Microorganisms are found in diverse communities whose structure and function are determined by interspecific interactions. Just as single species seldom exist in isolation, communities as a whole are also constantly challenged and affected by external species. Though much work has been done on characterizing how individual species affect each other through pairwise interactions, the joint effects of multiple species on a single (focal) species, remain under explored. As such, it is still unclear how single species effects combine to a community-level effect on a species of interest. To explore this relationship, we assayed thousands of communities of two, three, and four bacterial species, measuring the effect of single, pairs of, and trios of 61 affecting species on six different focal species. We found that when multiple species each have a negative effect on a focal species, their joint effect is typically not given by the sum of the effects of individual affecting species. Rather, they are dominated by the strongest individual-species effect. Therefore, while joint effects of multiple species are often non-additive, they can still be derived from the effects of individual species, making it plausible to map complex interaction networks based on pairwise measurements. This finding is important for understanding the fate of species introduced into an occupied environment, and is relevant for applications in medicine and agriculture, such as probiotics and biocontrol agents, as well as for ecological questions surrounding migrating and invasive species.

PMID:36852917 | DOI:10.7554/eLife.83398

Plant Physiol. 2023 Feb 28:kiad130. doi: 10.1093/plphys/kiad130. Online ahead of print.

ABSTRACT

Developmental programmed cell death (dPCD) controls a plethora of functions in plant growth and reproduction. In the root cap of Arabidopsis (Arabidopsis thaliana), dPCD functions to control organ size in balance with the continuous stem cell activity in the root meristem. Key regulators of root cap dPCD including SOMBRERO/ANAC033 (SMB) belong to the NAC family of transcription factors. Here we identify the C2H2 zinc finger protein ZINC FINGER OF ARABIDOPSIS THALIANA 14 ZAT14 as part of the gene regulatory network of root cap dPCD acting downstream of SMB. Similar to SMB, ZAT14 inducible misexpression leads to extensive ectopic cell death. Both the canonical EAR motif and a conserved L-box motif of ZAT14 act as transcriptional repression motifs and are required to trigger cell death. While a single zat14 mutant does not show a cell death-related phenotype, a quintuple mutant knocking out five related ZAT paralogs shows a delayed onset of dPCD execution in the columella and the adjacent lateral root cap. While ZAT14 is co-expressed with established dPCD-associated genes, it does not activate their expression. Our results suggest that ZAT14 acts as a transcriptional repressor controlling a so far uncharacterized sub-section of the dPCD gene regulatory network active in specific root cap tissues.

PMID:36852889 | DOI:10.1093/plphys/kiad130

Br J Haematol. 2023 Feb 28. doi: 10.1111/bjh.18708. Online ahead of print.

ABSTRACT

The tumour microenvironment (TME) plays a critical role in disease progression in multiple myeloma (MM). This study aimed to present an atlas of MM-TME in disease progression and explore TME-directed therapeutic strategies. We performed single-cell RNA sequencing (scRNAseq) in samples from different disease stages. We validated the findings by bulk RNAseq, flow cytometry (FCM) and in vitro and in vivo functional experiments. We delineated a compromised TME during disease progression, characterized by enrichment of exhausted NK cells and CD8+ T cells and reprogramming of macrophages (MPs). The reprogrammed tumour-associated MPs (TAMs) displayed a mixed phenotype showing both M1 and M2 features, with two TAM clusters exclusively present in the MM stage showing higher M2 scores. We validated the mixed M1/M2 phenotype in TAMs in a clinical cohort and verified phagocytic dysfunction in reprogrammed TAMs. Cellular interaction analysis identified two enriched ligand-receptor pairs between MPs and malignant plasma cells (PCs), including the SIRPA-CD47 pathway suppressing phagocytosis and the CD74-MIF (macrophage inhibitory factor) reshaping the phenotype of MPs. The expression of CD47 and MIF correlated with disease progression and adverse outcomes. We designed a dual-MP-targeted strategy by combining an anti-CD47 antibody and MIF inhibitor to activate phagocytosis and repolarize MP to a functional phenotype and proved its potent antitumour effect in vitro and in vivo. We drafted alterations in MM-TME during disease progression and unravelled TAM's reprogramming. The dual MP-targeted approach blocking both CD47 and MIF showed potent antitumour effects.

PMID:36852636 | DOI:10.1111/bjh.18708

Comput Struct Biotechnol J. 2023 Feb 8;21:1523-1532. doi: 10.1016/j.csbj.2023.02.006. eCollection 2023.

ABSTRACT

Regulatory networks structure and signaling pathways dynamics are uncovered in time- and resource consuming experimental work. However, it is increasingly supported by modeling, analytical and computational techniques as well as discrete mathematics and artificial intelligence applied to to extract knowledge from existing databases. This review is focused on mathematical modeling used to analyze dynamics and robustness of these networks. This paper presents a review of selected modeling methods that facilitate advances in molecular biology.

PMID:36851915 | PMC:PMC9958294 | DOI:10.1016/j.csbj.2023.02.006

Viruses. 2023 Jan 18;15(2):273. doi: 10.3390/v15020273.

ABSTRACT

There was an error in the original publication [...].

PMID:36851804 | DOI:10.3390/v15020273

Viruses. 2023 Feb 13;15(2):519. doi: 10.3390/v15020519.

ABSTRACT

Profile hidden Markov models (HMMs) are a powerful way of modeling biological sequence diversity and constitute a very sensitive approach to detecting divergent sequences. Here, we report the development of protocols for the rational design of profile HMMs. These methods were implemented on TABAJARA, a program that can be used to either detect all biological sequences of a group or discriminate specific groups of sequences. By calculating position-specific information scores along a multiple sequence alignment, TABAJARA automatically identifies the most informative sequence motifs and uses them to construct profile HMMs. As a proof-of-principle, we applied TABAJARA to generate profile HMMs for the detection and classification of two viral groups presenting different evolutionary rates: bacteriophages of the Microviridae family and viruses of the Flavivirus genus. We obtained conserved models for the generic detection of any Microviridae or Flavivirus sequence, and profile HMMs that can specifically discriminate Microviridae subfamilies or Flavivirus species. In another application, we constructed Cas1 endonuclease-derived profile HMMs that can discriminate CRISPRs and casposons, two evolutionarily related transposable elements. We believe that the protocols described here, and implemented on TABAJARA, constitute a generic toolbox for generating profile HMMs for the highly sensitive and specific detection of sequence classes.

PMID:36851733 | DOI:10.3390/v15020519

Viruses. 2023 Feb 1;15(2):414. doi: 10.3390/v15020414.

ABSTRACT

The prevalence and distribution of African alphaviruses such as chikungunya have increased in recent years. Therefore, a better understanding of the local distribution of alphaviruses in vectors across the African continent is important. Here, entomological surveillance was performed from 2014 to 2018 at selected sites in north-eastern parts of South Africa where alphaviruses have been identified during outbreaks in humans and animals in the past. Mosquitoes were collected using a net, CDC-light, and BG-traps. An alphavirus genus-specific nested RT-PCR was used for screening, and positive pools were confirmed by sequencing and phylogenetic analysis. We collected 64,603 mosquitoes from 11 genera, of which 39,035 females were tested. Overall, 1462 mosquito pools were tested, of which 21 were positive for alphaviruses. Sindbis (61.9%, N = 13) and Middelburg (28.6%, N = 6) viruses were the most prevalent. Ndumu virus was detected in two pools (9.5%, N = 2). No chikungunya positive pools were identified. Arboviral activity was concentrated in peri-urban, rural, and conservation areas. A range of Culicidae species, including Culex univittatus, Cx. pipiens s.l., Aedes durbanensis, and the Ae. dentatus group, were identified as potential vectors. These findings confirm the active circulation and distribution of alphaviruses in regions where human or animal infections were identified in South Africa.

PMID:36851627 | DOI:10.3390/v15020414

Viruses. 2023 Jan 28;15(2):372. doi: 10.3390/v15020372.

ABSTRACT

Noroviruses infect a wide range of mammals and are the major cause of gastroenteritis in humans. Recombination at the junction of ORF1 encoding nonstructural proteins and ORF2 encoding major capsid protein VP1 is a well-known feature of noroviruses. Using all available complete norovirus sequences, we systematically analyzed patterns of natural recombination in the genus Norovirus both throughout the genome and across the genogroups. Recombination events between nonstructural (ORF1) and structural genomic regions (ORF2 and ORF3) were found in all analyzed genogroups of noroviruses, although recombination was most prominent between members of GII, the most common genogroup that infects humans. The half-life times of recombinant forms (clades without evidence of recombination) of human GI and GII noroviruses were 10.4 and 8.4-11.3 years, respectively. There was evidence of many recent recombination events, and most noroviruses that differed by more than 18% of nucleotide sequence were recombinant relative to each other. However, there were no distinct recombination events between viruses that differed by over 42% in ORF2/3, consistent with the absence of systematic recombination between different genogroups. The few inter-genogroup recombination events most likely occurred between ancient viruses before they diverged into contemporary genogroups. The recombination events within ORF1 or between ORF2/3 were generally rare. Thus, noroviruses routinely exchange full structural and nonstructural blocks of the genome, providing a modular evolution.

PMID:36851586 | DOI:10.3390/v15020372

Vaccines (Basel). 2023 Feb 6;11(2):371. doi: 10.3390/vaccines11020371.

ABSTRACT

Most neutralizing antibodies neutralize the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by directly blocking the interactions between the spike glycoprotein receptor-binding domain (RBD) and its receptor, human angiotensin-converting enzyme 2 (ACE2). Here, we report a novel nanobody (Nb) identified by an RBD-ACE2 competitive panning method that could specifically bind to the RBD of SARS-CoV-2 with a high affinity (EC50 = 0.03 nM) and successfully block the binding between the RBD and ACE2 recombinant protein. A structural simulation of the RBD-VHH complex also supports a mechanism of the Nb to block the interaction between the RBD and ACE2. A pseudovirus assay of the Nb showed it could neutralize the WT pseudovirus with high potency (IC50 = 0.026 μg/mL). Furthermore, we measured its binding to phages displaying RBDs of different SARS-CoV-2 variants and found that it could bind to recombinant phages displaying the RBD of beta and delta variants. This study also provides a method of phage library competitive panning, which could be useful for directly screening high-affinity antibodies targeting important functional regions.

PMID:36851249 | DOI:10.3390/vaccines11020371

Sensors (Basel). 2023 Feb 6;23(4):1820. doi: 10.3390/s23041820.

ABSTRACT

The detection of beneficial microbes living within perennial ryegrass seed causing no apparent defects is challenging, even with the most sensitive and conventional methods, such as DNA genotyping. Using a near-infrared hyperspectral imaging system (NIR-HSI), we were able to discriminate not only the presence of the commercial NEA12 fungal endophyte strain but perennial ryegrass cultivars of diverse seed age and batch. A total of 288 wavebands were extracted for individual seeds from hyperspectral images. The optimal pre-processing methods investigated yielded the best partial least squares discriminant analysis (PLS-DA) classification model to discriminate NEA12 and without endophyte (WE) perennial ryegrass seed with a classification accuracy of 89%. Effective wavelength (EW) selection based on GA-PLS-DA resulted in the selection of 75 wavebands yielding 88.3% discrimination accuracy using PLS-DA. For cultivar identification, the artificial neural network discriminant analysis (ANN-DA) was the best-performing classification model, resulting in >90% classification accuracy for Trojan, Alto, Rohan, Governor and Bronsyn. EW selection using GA-PLS-DA resulted in 87 wavebands, and the PLS-DA model performed the best, with no extensive compromise in performance, resulting in >89.1% accuracy. The study demonstrates the use of NIR-HSI reflectance data to discriminate, for the first time, an associated beneficial fungal endophyte and five cultivars of perennial ryegrass seed, irrespective of seed age and batch. Furthermore, the negligible effects on the classification errors using EW selection improve the capability and deployment of optimized methods for real-time analysis, such as the use of low-cost multispectral sensors for single seed analysis and automated seed sorting devices.

PMID:36850417 | DOI:10.3390/s23041820

Nat Metab. 2023 Feb;5(2):294-313. doi: 10.1038/s42255-023-00741-x. Epub 2023 Feb 27.

ABSTRACT

Many cell biological and biochemical mechanisms controlling the fundamental process of eukaryotic cell division have been identified; however, the temporal dynamics of biosynthetic processes during the cell division cycle are still elusive. Here, we show that key biosynthetic processes are temporally segregated along the cell cycle. Using budding yeast as a model and single-cell methods to dynamically measure metabolic activity, we observe two peaks in protein synthesis, in the G1 and S/G2/M phase, whereas lipid and polysaccharide synthesis peaks only once, during the S/G2/M phase. Integrating the inferred biosynthetic rates into a thermodynamic-stoichiometric metabolic model, we find that this temporal segregation in biosynthetic processes causes flux changes in primary metabolism, with an acceleration of glucose-uptake flux in G1 and phase-shifted oscillations of oxygen and carbon dioxide exchanges. Through experimental validation of the model predictions, we demonstrate that primary metabolism oscillates with cell-cycle periodicity to satisfy the changing demands of biosynthetic processes exhibiting unexpected dynamics during the cell cycle.

PMID:36849832 | DOI:10.1038/s42255-023-00741-x

Nat Plants. 2023 Feb 27. doi: 10.1038/s41477-023-01358-4. Online ahead of print.

ABSTRACT

Rhamnogalacturonan I (RGI) is a structurally complex pectic polysaccharide with a backbone of alternating rhamnose and galacturonic acid residues substituted with arabinan and galactan side chains. Galactan synthase 1 (GalS1) transfers galactose and arabinose to either extend or cap the β-1,4-galactan side chains of RGI, respectively. Here we report the structure of GalS1 from Populus trichocarpa, showing a modular protein consisting of an N-terminal domain that represents the founding member of a new family of carbohydrate-binding module, CBM95, and a C-terminal glycosyltransferase family 92 (GT92) catalytic domain that adopts a GT-A fold. GalS1 exists as a dimer in vitro, with stem domains interacting across the chains in a 'handshake' orientation that is essential for maintaining stability and activity. In addition to understanding the enzymatic mechanism of GalS1, we gained insight into the donor and acceptor substrate binding sites using deep evolutionary analysis, molecular simulations and biochemical studies. Combining all the results, a mechanism for GalS1 catalysis and a new model for pectic galactan side-chain addition are proposed.

PMID:36849618 | DOI:10.1038/s41477-023-01358-4

Comput Struct Biotechnol J. 2023 Feb 27;21:2446-2453. doi: 10.1016/j.csbj.2023.02.047. eCollection 2023.

ABSTRACT

Peptide retention time (RT) prediction algorithms are tools to study and identify the physicochemical properties that drive the peptide-sorbent interaction. Traditional RT algorithms use multiple linear regression with manually curated parameters to determine the degree of direct contribution for each parameter and improvements to RT prediction accuracies relied on superior feature engineering. Deep learning led to a significant increase in RT prediction accuracy and automated feature engineering via chaining multiple learning modules. However, the significance and the identity of these extracted variables are not well understood due to the inherent complexity when interpreting "relationships-of-relationships" found in deep learning variables. To achieve both accuracy and interpretability simultaneously, we isolated individual modules used in deep learning and the isolated modules are the shallow learners employed for RT prediction in this work. Using a shallow convolutional neural network (CNN) and gated recurrent unit (GRU), we find that the spatial features obtained via the CNN correlate with real-world physicochemical properties namely cross-collisional sections (CCS) and variations of assessable surface area (ASA). Furthermore, we determined that the discovered parameters are "micro-coefficients" that contribute to the "macro-coefficient" - hydrophobicity. Manually embedding CCS and the variations of ASA to the GRU model yielded an R2 = 0.981 using only 525 variables and can represent 88% of the ∼110,000 tryptic peptides used in our dataset. This work highlights the feature discovery process of our shallow learners can achieve beyond traditional RT models in performance and have better interpretability when compared with the deep learning RT algorithms found in the literature.

PMID:37090433 | PMC:PMC10113922 | DOI:10.1016/j.csbj.2023.02.047

Cell Rep. 2023 Feb 26;42(3):112153. doi: 10.1016/j.celrep.2023.112153. Online ahead of print.

ABSTRACT

Pyruvate dehydrogenase (PDH) is the central enzyme connecting glycolysis and the tricarboxylic acid (TCA) cycle. The importance of PDH function in T helper 17 (Th17) cells still remains to be studied. Here, we show that PDH is essential for the generation of a glucose-derived citrate pool needed for Th17 cell proliferation, survival, and effector function. In vivo, mice harboring a T cell-specific deletion of PDH are less susceptible to developing experimental autoimmune encephalomyelitis. Mechanistically, the absence of PDH in Th17 cells increases glutaminolysis, glycolysis, and lipid uptake in a mammalian target of rapamycin (mTOR)-dependent manner. However, cellular citrate remains critically low in mutant Th17 cells, which interferes with oxidative phosphorylation (OXPHOS), lipid synthesis, and histone acetylation, crucial for transcription of Th17 signature genes. Increasing cellular citrate in PDH-deficient Th17 cells restores their metabolism and function, identifying a metabolic feedback loop within the central carbon metabolism that may offer possibilities for therapeutically targeting Th17 cell-driven autoimmunity.

PMID:36848289 | DOI:10.1016/j.celrep.2023.112153

Cell Tissue Res. 2023 Feb 27. doi: 10.1007/s00441-023-03744-y. Online ahead of print.

ABSTRACT

Asthenozoospermia, characterized by low sperm motility, is one of the most common causes of male infertility. While many intrinsic and extrinsic factors are involved in the etiology of asthenozoospermia, the molecular basis of this condition remains unclear. Since sperm motility results from a complex flagellar structure, an in-depth proteomic analysis of the sperm tail can uncover mechanisms underlying asthenozoospermia. This study quantified the proteomic profile of 40 asthenozoospermic sperm tails and 40 controls using TMT-LC-MS/MS. Overall, 2140 proteins were identified and quantified where 156 proteins have not been described earlier in sperm tail. There were 409 differentially expressed proteins (250 upregulated and 159 downregulated) in asthenozoospermia which by far is the highest number reported earlier. Further, bioinformatics analysis revealed several biological processes, including mitochondrial-related energy production, oxidative phosphorylation (OXPHOS), citric acid cycle (CAC), cytoskeleton, stress response, and protein metabolism altered in asthenozoospermic sperm tail samples. Collectively, our findings reveal the importance of mitochondrial energy production and induced stress response as potential mechanisms involved in the loss of sperm motility in asthenozoospermia.

PMID:36847810 | DOI:10.1007/s00441-023-03744-y

Microbiol Spectr. 2023 Feb 27:e0382622. doi: 10.1128/spectrum.03826-22. Online ahead of print.

ABSTRACT

The genus Yersinia includes a large variety of nonpathogenic and life-threatening pathogenic bacteria, which cause a broad spectrum of diseases in humans and animals, such as plague, enteritis, Far East scarlet-like fever (FESLF), and enteric redmouth disease. Like most clinically relevant microorganisms, Yersinia spp. are currently subjected to intense multi-omics investigations whose numbers have increased extensively in recent years, generating massive amounts of data useful for diagnostic and therapeutic developments. The lack of a simple and centralized way to exploit these data led us to design Yersiniomics, a web-based platform allowing straightforward analysis of Yersinia omics data. Yersiniomics contains a curated multi-omics database at its core, gathering 200 genomic, 317 transcriptomic, and 62 proteomic data sets for Yersinia species. It integrates genomic, transcriptomic, and proteomic browsers, a genome viewer, and a heatmap viewer to navigate within genomes and experimental conditions. For streamlined access to structural and functional properties, it directly links each gene to GenBank, the Kyoto Encyclopedia of Genes and Genomes (KEGG), UniProt, InterPro, IntAct, and the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) and each experiment to Gene Expression Omnibus (GEO), the European Nucleotide Archive (ENA), or the Proteomics Identifications Database (PRIDE). Yersiniomics provides a powerful tool for microbiologists to assist with investigations ranging from specific gene studies to systems biology studies. IMPORTANCE The expanding genus Yersinia is composed of multiple nonpathogenic species and a few pathogenic species, including the deadly etiologic agent of plague, Yersinia pestis. In 2 decades, the number of genomic, transcriptomic, and proteomic studies on Yersinia grew massively, delivering a wealth of data. We developed Yersiniomics, an interactive web-based platform, to centralize and analyze omics data sets on Yersinia species. The platform allows user-friendly navigation between genomic data, expression data, and experimental conditions. Yersiniomics will be a valuable tool to microbiologists.

PMID:36847572 | DOI:10.1128/spectrum.03826-22

Bioinformatics. 2023 Feb 27:btad104. doi: 10.1093/bioinformatics/btad104. Online ahead of print.

ABSTRACT

SUMMARY: Leveraging local ancestry and haplotype information in genome-wide association studies and downstream analyses can improve the utility of genomics for individuals from diverse and recently admixed ancestries. However, most existing simulation, visualization, and variant analysis frameworks are based on variant-level analysis and do not automatically handle these features. We present haptools, an open-source toolkit for performing local-ancestry aware and haplotype-based analysis of complex traits. Haptools supports fast simulation of admixed genomes, visualization of admixture tracks, simulation of haplotype- and local ancestry-specific phenotype effects, and a variety of file operations and statistics computed in a haplotype-aware manner.

AVAILABILITY: Haptools is freely available at https://github.com/cast-genomics/haptools.

DOCUMENTATION: Detailed documentation is available at https://haptools.readthedocs.io.

SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

PMID:36847450 | DOI:10.1093/bioinformatics/btad104

Elife. 2023 Feb 27;12:e79408. doi: 10.7554/eLife.79408. Online ahead of print.

ABSTRACT

The shuffling of genetic material facilitated by meiotic crossovers is a critical driver of genetic variation. Therefore, the number and positions of crossover events must be carefully controlled. In Arabidopsis, an obligate crossover and repression of nearby crossovers on each chromosome pair are abolished in mutants that lack the synaptonemal complex (SC), a conserved protein scaffold. We use mathematical modelling and quantitative super-resolution microscopy to explore and mechanistically explain meiotic crossover pattering in Arabidopsis lines with full, incomplete or abolished synapsis. For zyp1 mutants, which lack an SC, we develop a coarsening model in which crossover precursors globally compete for a limited pool of the pro-crossover factor HEI10, with dynamic HEI10 exchange mediated through the nucleoplasm. We demonstrate that this model is capable of quantitatively reproducing and predicting zyp1 experimental crossover patterning and HEI10 foci intensity data. Additionally, we find that a model combining both SC- and nucleoplasm-mediated coarsening can explain crossover patterning in wild-type Arabidopsis and in pch2 mutants, which display partial synapsis. Together, our results reveal that regulation of crossover patterning in wild-type Arabidopsis and SC defective mutants likely act through the same underlying coarsening mechanism, differing only in the spatial compartments through which the pro-crossover factor diffuses.

PMID:36847348 | DOI:10.7554/eLife.79408

Elife. 2023 Feb 27;12:e80942. doi: 10.7554/eLife.80942. Online ahead of print.

ABSTRACT

Predicting the function of a protein from its amino acid sequence is a long-standing challenge in bioinformatics. Traditional approaches use sequence alignment to compare a query sequence either to thousands of models of protein families or to large databases of individual protein sequences. Here we introduce ProteInfer, which instead employs deep convolutional neural networks to directly predict a variety of protein functions - EC numbers and GO terms - directly from an unaligned amino acid sequence. This approach provides precise predictions which complement alignment-based methods, and the computational efficiency of a single neural network permits novel and lightweight software interfaces, which we demonstrate with an in-browser graphical interface for protein function prediction in which all computation is performed on the user's personal computer with no data uploaded to remote servers. Moreover, these models place full-length amino acid sequences into a generalised functional space, facilitating downstream analysis and interpretation. To read the interactive version of this paper, please visit https://google-research.github.io/proteinfer/.

PMID:36847334 | DOI:10.7554/eLife.80942