Beta turn propensity and a model polymer scaling exponent identify intrinsically disordered phase-separating proteins

PSTP: accurate residue-level phase separation prediction using protein conformational and language model embeddings

Phase separation (PS) is essential in cellular processes and disease mechanisms, highlighting the need for predictive algorithms to analyze uncharacterized sequences and accelerate experimental validation. Current high-accuracy methods often rely on extensive annotations or handcrafted features, limiting their generalizability to sequences lacking such annotations and making it difficult to identify key protein regions involved in PS. We introduce Phase Separation's Transfer-learning Prediction (PSTP), which combines conformational embeddings with large language model embeddings, enabling state-of-the-art PS predictions from protein sequences alone. PSTP performs well across various prediction scenarios and shows potential for predicting novel-designed artificial proteins. Additionally, PSTP provides residue-level predictions that are highly correlated with experimentally validated PS regions. By analyzing 160 000+ variants, PSTP characterizes the strong link between the incidence of pathogenic variants and residue-level PS propensities in unconserved intrinsically disordered regions, offering insights into underexplored mutation effects. PSTP's sliding-window optimization reduces its memory usage to a few hundred megabytes, facilitating rapid execution on typical CPUs and GPUs. Offered via both a web server and an installable Python package, PSTP provides a versatile tool for decoding protein PS behavior and supporting disease-focused research.

The Myotubularin Related Proteins and the Untapped Interaction Potential of Their Disordered C-Terminal Regions

Intrinsically disordered regions (IDRs) of proteins remain understudied with enigmatic sequence features relevant to their functions. Members of the myotubularin-related protein (MTMR) family contain uncharacterized IDRs. After decades of research on their phosphatase activity, recent work on the C-terminal IDRs of MTMR7 revealed new interactions and important new functions beyond the phosphatase function. Here we take a broader look at the C-terminal domains (CTDs) of 14 human MTMRs and use bioinformatic tools and biophysical methods to ask which other functions may be probable in this protein family. The predictions show that the CTDs are disordered and carry short linear motifs (SLiMs) important for targeting of MTMRs to defined subcellular compartments and implicating them in signaling, phase separation, interaction with diverse proteins, including transcription factors and are of relevance for cancer research and neuroscience. We also present experimental methods to study the CTDs and use them to characterize the coiled coil (CC) domains of MTMR7 and MTMR9. We show homo- and hetero-oligomerization with preference for MTMR7-CC to form dimers, while MTMR9-CC forms trimers. We relate the results to sequence features and make predictions for the structural landscape of other MTMRs. Our work gives a broad insight into the so far unrecognized features and SLiMs in MTMR-CTDs, and provides the basis for more in-depth experimental research on this diverse protein family and understudied IDRs in proteins in general.

ExploreTurns: A web tool for the exploration, analysis, and classification of beta turns and structured loops in proteins; application to beta-bulge and Schellman loops, Asx helix caps, beta hairpins, and other hydrogen-bonded motifs

The most common type of protein secondary structure after the alpha helix and beta sheet is the four-residue beta turn, which plays many key structural and functional roles. Existing tools for the study of beta turns operate in backbone dihedral-angle (Ramachandran) space, which presents challenges for the visualization, comparison and analysis of the wide range of turn conformations. In this work, a new turn-local coordinate system and structural alignment, together with a set of geometric descriptors for turn backbone shape, are incorporated into ExploreTurns, a web facility for the exploration, analysis, geometric tuning and retrieval of beta turns and their contexts which combines the advantages of Ramachandran- and Euclidean-space representations. Due to the prevalence of beta turns in proteins, this facility, supported by its interpreter for a new general nomenclature which classifies H-bonded loop motifs and beta hairpins, serves as an exploratory browser and analysis tool for most loop structure. The tool is applied to the detection of new H-bonded loops, including short and "double" Schellman loops, a large family of beta-bulge loops with a range of geometries and H-bond topologies, and other motifs. Other applications presented here include the mapping of sequence preferences in Asx helix N-caps and an investigation of the depth dependence of beta-turn geometry. ExploreTurns, available at www.betaturn.com, should prove useful in research, education, and applications such as protein design, in which an enhanced Euclidean-space picture of turn and motif structure and the ability to identify and tune structures suited to particular requirements may improve performance.

Molecular crowding and amyloidogenic self-assembly: Emergent perspectives from modern computations

In recent decades, the conventional protein folding paradigm has been challenged by intriguing properties of disordered peptide sequences that do not adopt stably folded conformations. Such intrinsically disordered proteins and protein regions (IDPs and IDRs) are poised uniquely in biology due to their propensity for self-aggregation, amyloidogenesis, and correlations with a cluster of debilitating diseases. Complexities underlying their structural and functional manifestations are enhanced in the presence of molecular crowding via non-specific protein-protein and protein-solvent contacts. Enabled by technological advances, physics-based algorithms, and data science, modern computer simulations provide unprecedented insights into the structure, function, dynamics, and thermodynamics of complex macromolecular systems. These characteristics are frequently correlated and manifest into unique observables. This chapter presents an overview of how such methodologies can lend insights and drive investigations into the molecular trifecta of crowding, protein self-aggregation, and amyloidogenesis. It begins with a general overview of disordered proteins in relation to biological function and of a suite of relevant experimental methods. Specific examples are showcased in the biological context. This is followed by a description of the computational approaches that supplant experimental efforts, with an elaboration on enhanced molecular simulation methods. The chapter concludes by alluding to expanded possibilities in disease amelioration.

Expression of spider silk protein in tobacco improves drought tolerance with minimal effects on its mechanotype

Spider silk, especially dragline silk from golden silk spiders (Trichonephila clavipes), is an excellent natural material with remarkable mechanical properties. Many studies have focused on the use of plants as biofactories for the production of recombinant spider silk. However, the effects of this material on the mechanical properties or physiology of transgenic plants remain poorly understood. Since glycine-rich proteins play key roles in plants, we evaluated the effects of a glycine-rich spider silk protein on plant mechanical properties (mechanotype) and physiology. We generated tobacco (Nicotiana tabacum) plants producing a nucleus- or plastid-encoded partial component of dragline silk, MaSp1 (major ampullate spidroin-1; MaSp1-tobacco), containing six repetitive glycine-rich and polyalanine tandem domains. MaSp1 accumulation had minimal effect on leaf mechanical properties, but improved drought tolerance. Transcriptome analysis of drought-stressed MaSp1-tobacco revealed the upregulation of genes involved in stress response, antioxidant activity, cellular metabolism and homeostasis, and phenylpropanoid biosynthesis. The effects of drought treatment differed between the nucleus- and the plastid-encoded MaSp1-tobacco, with the latter showing a stronger transcriptomic response and a higher total antioxidant status (TAS). Well-watered MaSp1-tobacco displayed elevated levels of the stress phytohormone ABA, leading to stomatal closure, reduced water loss, activation of stress response, and increased TAS. We show that the moderately enhanced ABA content in these plants plays a pivotal role in drought tolerance, alongside, ABA priming, which causes overall adjustments in multiple drought tolerance mechanisms. Thus, our findings highlight the potential of utilizing glycine-rich spider silk proteins to enhance plant resilience to drought.

Yeast Glucan Remodeling Protein Bgl2p: Amyloid Properties and the Mode of Attachment in Cell Wall

Bgl2p is a major, conservative, constitutive glucanosyltransglycosylase of the yeast cell wall (CW) with amyloid amino acid sequences, strongly non-covalently anchored in CW, but is able to leave it. In the environment, Bgl2p can form fibrils and/or participate in biofilm formation. Despite a long study, the question of how Bgl2p is anchored in CW remains unclear. Earlier, it was demonstrated that Bgl2p lost the ability to attach in CW and to fibrillate after the deletion of nine amino acids in its C-terminal region (CTR). Here, we demonstrated that a Bgl2p anchoring is weakened by substitution Glu-233/Ala in the active center. Using AlphaFold and molecular modeling approach, we demonstrated the role of CTR on Bgl2p attachment and supposed the conformational possibilities determined by the presence or absence of an intramolecular disulfide bond, forming by Cys-310, leading to accessibility of amyloid sequence and β-turns localized in CTR of Bgl2p for protein interactions. We hypothesized the mode of Bgl2p attachment in CW. Using atomic force microscopy, we investigated fibrillar structures formed by peptide V187MANAFSYWQ196 and suggested that it can serve as a factor leading to the induction of amyloid formation during interaction of Bgl2p with other proteins and is of medical interest being located close to the surface of the molecule.

RNA-dependent RNA polymerase of predominant human norovirus forms liquid-liquid phase condensates as viral replication factories

Many viral proteins form biomolecular condensates via liquid-liquid phase separation (LLPS) to support viral replication and evade host antiviral responses, and thus, they are potential targets for designing antivirals. In the case of nonenveloped positive-sense RNA viruses, forming such condensates for viral replication is unclear and less understood. Human noroviruses (HuNoVs) are positive-sense RNA viruses that cause epidemic and sporadic gastroenteritis worldwide. Here, we show that the RNA-dependent RNA polymerase (RdRp) of pandemic GII.4 HuNoV forms distinct condensates that exhibit all the signature properties of LLPS with sustained polymerase activity and the capability of recruiting components essential for viral replication. We show that such condensates are formed in HuNoV-infected human intestinal enteroid cultures and are the sites for genome replication. Our studies demonstrate the formation of phase-separated condensates as replication factories in a positive-sense RNA virus, which plausibly is an effective mechanism to dynamically isolate RdRp replicating the genomic RNA from interfering with the ribosomal translation of the same RNA.

RNA-dependent RNA polymerase of predominant human norovirus forms liquid-liquid phase condensates as viral replication factories

Many viral proteins form biomolecular condensates via liquid-liquid phase separation (LLPS) to support viral replication and evade host antiviral responses, and thus, they are potential targets for designing antivirals. In the case of non-enveloped positive-sense RNA viruses, forming such condensates for viral replication is unclear and less understood. Human noroviruses (HuNoV) are positive-sense RNA viruses that cause epidemic and sporadic gastroenteritis worldwide. Here, we show that the RNA-dependent-RNA polymerase (RdRp) of pandemic GII.4 HuNoV forms distinct condensates that exhibit all the signature properties of LLPS with sustained polymerase activity and the capability of recruiting components essential for viral replication. We show that such condensates are formed in HuNoV-infected human intestinal enteroid cultures and are the sites for genome replication. Our studies demonstrate the formation of phase separated condensates as replication factories in a positive-sense RNA virus, which plausibly is an effective mechanism to dynamically isolate RdRp replicating the genomic RNA from interfering with the ribosomal translation of the same RNA.

Biophysical characterization of high-confidence, small human proteins

Significant efforts have been made to characterize the biophysical properties of proteins. Small proteins have received less attention because their annotation has historically been less reliable. However, recent improvements in sequencing, proteomics, and bioinformatics techniques have led to the high-confidence annotation of small open reading frames (smORFs) that encode for functional proteins, producing smORF-encoded proteins (SEPs). SEPs have been found to perform critical functions in several species, including humans. While significant efforts have been made to annotate SEPs, less attention has been given to the biophysical properties of these proteins. We characterized the distributions of predicted and curated biophysical properties, including sequence composition, structure, localization, function, and disease association of a conservative list of previously identified human SEPs. We found significant differences between SEPs and both larger proteins and control sets. In addition, we provide an example of how our characterization of biophysical properties can contribute to distinguishing protein-coding smORFs from noncoding ones in otherwise ambiguous cases.

VUS next in rare diseases? Deciphering genetic determinants of biomolecular condensation

The diagnostic odysseys for rare disease patients are getting shorter as next-generation sequencing becomes more widespread. However, the complex genetic diversity and factors influencing expressivity continue to challenge accurate diagnosis, leaving more than 50% of genetic variants categorized as variants of uncertain significance.Genomic expression intricately hinges on localized interactions among its products. Conventional variant prioritization, biased towards known disease genes and the structure-function paradigm, overlooks the potential impact of variants shaping the composition, location, size, and properties of biomolecular condensates, genuine membraneless organelles swiftly sensing and responding to environmental changes, and modulating expressivity.To address this complexity, we propose to focus on the nexus of genetic variants within biomolecular condensates determinants. Scrutinizing variant effects in these membraneless organelles could refine prioritization, enhance diagnostics, and unveil the molecular underpinnings of rare diseases. Integrating comprehensive genome sequencing, transcriptomics, and computational models can unravel variant pathogenicity and disease mechanisms, enabling precision medicine. This paper presents the rationale driving our proposal and describes a protocol to implement this approach. By fusing state-of-the-art knowledge and methodologies into the clinical practice, we aim to redefine rare diseases diagnosis, leveraging the power of scientific advancement for more informed medical decisions.

Recurrent evolution of adhesive defence systems in amphibians by parallel shifts in gene expression

Natural selection can drive organisms to strikingly similar adaptive solutions, but the underlying molecular mechanisms often remain unknown. Several amphibians have independently evolved highly adhesive skin secretions (glues) that support a highly effective antipredator defence mechanism. Here we demonstrate that the glue of the Madagascan tomato frog, Dyscophus guineti, relies on two interacting proteins: a highly derived member of a widespread glycoprotein family and a galectin. Identification of homologous proteins in other amphibians reveals that these proteins attained a function in skin long before glues evolved. Yet, major elevations in their expression, besides structural changes in the glycoprotein (increasing its structural disorder and glycosylation), caused the independent rise of glues in at least two frog lineages. Besides providing a model for the chemical functioning of animal adhesive secretions, our findings highlight how recruiting ancient molecular templates may facilitate the recurrent evolution of functional innovations.

Biophysical characterization of high-confidence, small human proteins

Significant efforts have been made to characterize the biophysical properties of proteins. Small proteins have received less attention because their annotation has historically been less reliable. However, recent improvements in sequencing, proteomics, and bioinformatics techniques have led to the high-confidence annotation of small open reading frames (smORFs) that encode for functional proteins, producing smORF-encoded proteins (SEPs). SEPs have been found to perform critical functions in several species, including humans. While significant efforts have been made to annotate SEPs, less attention has been given to the biophysical properties of these proteins. We characterized the distributions of predicted and curated biophysical properties, including sequence composition, structure, localization, function, and disease association of a conservative list of previously identified human SEPs. We found significant differences between SEPs and both larger proteins and control sets. Additionally, we provide an example of how our characterization of biophysical properties can contribute to distinguishing protein-coding smORFs from non-coding ones in otherwise ambiguous cases.

Protein-Functionalized Gold Nanospheres with Tunable Photothermal Efficiency for the Near-Infrared Photothermal Ablation of Biofilms

Temperature-responsive nanostructures with high antimicrobial efficacy are attractive for therapeutic applications against multidrug-resistant bacteria. Here, we report temperature-responsive nanospheres (TRNs) engineered to undergo self-association and agglomeration above a tunable transition temperature (Tt). The temperature-responsive behavior of the nanoparticles is obtained by functionalizing citrate-capped spherical gold nanoparticles (AuNPs) with elastin-like polypeptides (ELPs). Using protein design principles, we achieve a broad range of attainable Tt values and photothermal conversion efficiencies (η). Two approaches were used to adjust this range: First, by altering the position of the cysteine residue used to attach ELP to the AuNP, we attained a Tt range from 34 to 42 °C. Then, by functionalizing the AuNP with an additional small globular protein, we could extend this range to 34-50 °C. Under near-infrared (NIR) light exposure, all TRNs exhibited reversible agglomeration. Moreover, they showed an enhanced photothermal conversion efficiency in their agglomerated state relative to the dispersed state. Despite their spherical shape, TRNs have a photothermal conversion efficiency approaching that of gold nanorods (η = 68 ± 6%), yet unlike nanorods, the synthesis of TRNs requires no cytotoxic compounds. Finally, we tested TRNs for the photothermal ablation of biofilms. Above Tt, NIR irradiation of TRNs resulted in a 10,000-fold improvement in killing efficiency compared to untreated controls (p < 0.0001). Below Tt, no enhanced antibiofilm effect was observed. In conclusion, engineering the interactions between proteins and nanoparticles enables the tunable control of TRNs, resulting in a novel antibiofilm nanomaterial with low cytotoxicity.

Regulation of micro- and small-exon retention and other splicing processes by GRP20 for flower development

Pre-mRNA splicing is crucial for gene expression and depends on the spliceosome and splicing factors. Plant exons have an average size of ~180 nucleotides and typically contain motifs for interactions with spliceosome and splicing factors. Micro exons (<51 nucleotides) are found widely in eukaryotes and in genes for plant development and environmental responses. However, little is known about transcript-specific regulation of splicing in plants and about the regulators for micro exon splicing. Here we report that glycine-rich protein 20 (GRP20) is an RNA-binding protein and required for splicing of ~2,100 genes including those functioning in flower development and/or environmental responses. Specifically, GRP20 is required for micro-exon retention in transcripts of floral homeotic genes; these micro exons are conserved across angiosperms. GRP20 is also important for small-exon (51-100 nucleotides) splicing. In addition, GRP20 is required for flower development. Furthermore, GRP20 binds to poly-purine motifs in micro and small exons and a spliceosome component; both RNA binding and spliceosome interaction are important for flower development and micro-exon retention. Our results provide new insights into the mechanisms of micro-exon retention in flower development.

Copper drives prion protein phase separation and modulates aggregation

Prion diseases are characterized by prion protein (PrP) transmissible aggregation and neurodegeneration, which has been linked to oxidative stress. The physiological function of PrP seems related to sequestering of redox-active Cu2+, and Cu2+ dyshomeostasis is observed in prion disease brain. It is unclear whether Cu2+ contributes to PrP aggregation, recently shown to be mediated by PrP condensation. This study indicates that Cu2+ promotes PrP condensation in live cells at the cell surface and in vitro through copartitioning. Molecularly, Cu2+ inhibited PrP β-structure and hydrophobic residues exposure. Oxidation, induced by H2O2, triggered liquid-to-solid transition of PrP:Cu2+ condensates and promoted amyloid-like PrP aggregation. In cells, overexpression of PrPC initially protected against Cu2+ cytotoxicity but led to PrPC aggregation upon extended copper exposure. Our data suggest that PrP condensates function as a buffer for copper that prevents copper toxicity but can transition into PrP aggregation at prolonged oxidative stress.

ParSe 2.0: A web tool to identify drivers of protein phase separation at the proteome level

We have developed an algorithm, ParSe, which accurately identifies from the primary sequence those protein regions likely to exhibit physiological phase separation behavior. Originally, ParSe was designed to test the hypothesis that, for flexible proteins, phase separation potential is correlated to hydrodynamic size. While our results were consistent with that idea, we also found that many different descriptors could successfully differentiate between three classes of protein regions: folded, intrinsically disordered, and phase-separating intrinsically disordered. Consequently, numerous combinations of amino acid property scales can be used to make robust predictions of protein phase separation. Built from that finding, ParSe 2.0 uses an optimal set of property scales to predict domain-level organization and compute a sequence-based prediction of phase separation potential. The algorithm is fast enough to scan the whole of the human proteome in minutes on a single computer and is equally or more accurate than other published predictors in identifying proteins and regions within proteins that drive phase separation. Here, we describe a web application for ParSe 2.0 that may be accessed through a browser by visiting https://stevewhitten.github.io/Parse_v2_FASTA to quickly identify phase-separating proteins within large sequence sets, or by visiting https://stevewhitten.github.io/Parse_v2_web to evaluate individual protein sequences.

Near-Infrared Photothermal Ablation of Biofilms using ProteinFunctionalized Gold Nanospheres with a Tunable Temperature Response

Temperature-responsive nanostructures with high antimicrobial efficacy are attractive for therapeutic applications against multi-drug-resistant bacteria. Here, we report temperature-responsive nanospheres (TRNs) that are engineered to undergo self-association and agglomeration above a tunable transition temperature (Tt). Temperature-responsive behavior of the nanoparticles is obtained by functionalizing citrate-capped, spherical gold nanoparticles (AuNPs) with elastin-like polypeptides (ELPs). Using protein design principles, we achieve a broad range of attainable Tt values and photothermal conversion efficiencies (η). Two approaches were used to adjust this range: First, by altering the position of the cysteine residue used to attach ELP to the AuNP, we attained a Tt range from 34-42 °C. Then, functionalizing the AuNP with an additional small globular protein, we were able to extend this range to 34-50 °C. Under near-infrared (NIR) light exposure, all TRNs exhibited reversible agglomeration. Moreover, they showed enhanced photothermal conversion efficiency in their agglomerated state relative to the dispersed state. Despite their spherical shape, TRNs have a photothermal conversion efficiency approaching that of gold nanorods (η = 68±6%), yet unlike nanorods, the synthesis of TRNs requires no cytotoxic compounds. Finally, we tested TRNs for photothermal ablation of biofilms. Above Tt, NIR irradiation of TRNs resulted in a 10,000-fold improvement in killing efficiency compared to untreated controls (p < 0.0001). Below Tt, no enhanced anti-biofilm effect was observed. In conclusion, engineering the interactions between proteins and nanoparticles enables the tunable control of TRNs, resulting in a novel, anti-biofilm nanomaterial with low cytotoxicity.

Kinase regulation by liquid-liquid phase separation

Liquid-liquid phase separation (LLPS) is emerging as a mechanism of spatiotemporal regulation that could answer long-standing questions about how order is achieved in biochemical signaling. In this review we discuss how LLPS orchestrates kinase signaling, either by creating condensate structures that are sensed by kinases or by direct LLPS of kinases, cofactors, and substrates - thereby acting as a mechanism to compartmentalize kinase-substrate relationships, and in some cases also sequestering the kinase away from inhibitory factors. We also examine the possibility that selective pressure promotes genomic rearrangements that fuse pro-growth kinases to LLPS-prone protein sequences, which in turn drives aberrant kinase activation through LLPS.

G-Quadruplexes in Nuclear Biomolecular Condensates

G-quadruplexes (G4s) have long been implicated in the regulation of chromatin packaging and gene expression. These processes require or are accelerated by the separation of related proteins into liquid condensates on DNA/RNA matrices. While cytoplasmic G4s are acknowledged scaffolds of potentially pathogenic condensates, the possible contribution of G4s to phase transitions in the nucleus has only recently come to light. In this review, we summarize the growing evidence for the G4-dependent assembly of biomolecular condensates at telomeres and transcription initiation sites, as well as nucleoli, speckles, and paraspeckles. The limitations of the underlying assays and the remaining open questions are outlined. We also discuss the molecular basis for the apparent permissive role of G4s in the in vitro condensate assembly based on the interactome data. To highlight the prospects and risks of G4-targeting therapies with respect to the phase transitions, we also touch upon the reported effects of G4-stabilizing small molecules on nuclear biomolecular condensates.

Intrinsically disordered regions that drive phase separation form a robustly distinct protein class

Protein phase separation is thought to be a primary driving force for the formation of membrane-less organelles, which control a wide range of biological functions from stress response to ribosome biogenesis. Among phase-separating (PS) proteins, many have intrinsically disordered regions (IDRs) that are needed for phase separation to occur. Accurate identification of IDRs that drive phase separation is important for testing the underlying mechanisms of phase separation, identifying biological processes that rely on phase separation, and designing sequences that modulate phase separation. To identify IDRs that drive phase separation, we first curated datasets of folded, ID, and PS ID sequences. We then used these sequence sets to examine how broadly existing amino acid property scales can be used to distinguish between the three classes of protein regions. We found that there are robust property differences between the classes and, consequently, that numerous combinations of amino acid property scales can be used to make robust predictions of protein phase separation. This result indicates that multiple, redundant mechanisms contribute to the formation of phase-separated droplets from IDRs. The top-performing scales were used to further optimize our previously developed predictor of PS IDRs, ParSe. We then modified ParSe to account for interactions between amino acids and obtained reasonable predictive power for mutations that have been designed to test the role of amino acid interactions in driving protein phase separation. Collectively, our findings provide further insight into the classification of IDRs and the elements involved in protein phase separation.

An Interpretable Machine-Learning Algorithm to Predict Disordered Protein Phase Separation Based on Biophysical Interactions

Protein phase separation is increasingly understood to be an important mechanism of biological organization and biomaterial formation. Intrinsically disordered protein regions (IDRs) are often significant drivers of protein phase separation. A number of protein phase-separation-prediction algorithms are available, with many being specific for particular classes of proteins and others providing results that are not amenable to the interpretation of the contributing biophysical interactions. Here, we describe LLPhyScore, a new predictor of IDR-driven phase separation, based on a broad set of physical interactions or features. LLPhyScore uses sequence-based statistics from the RCSB PDB database of folded structures for these interactions, and is trained on a manually curated set of phase-separation-driving proteins with different negative training sets including the PDB and human proteome. Competitive training for a variety of physical chemical interactions shows the greatest contribution of solvent contacts, disorder, hydrogen bonds, pi–pi contacts, and kinked beta-structures to the score, with electrostatics, cation–pi contacts, and the absence of a helical secondary structure also contributing. LLPhyScore has strong phase-separation-prediction recall statistics and enables a breakdown of the contribution from each physical feature to a sequence’s phase-separation propensity, while recognizing the interdependence of many of these features. The tool should be a valuable resource for guiding experiments and providing hypotheses for protein function in normal and pathological states, as well as for understanding how specificity emerges in defining individual biomolecular condensates.