Cross-layer transmission realized by light-emitting memristor for constructing ultra-deep neural network with transfer learning ability.

Deep neural networks have revolutionized several domains, including autonomous driving, cancer detection, and drug design, and are the foundation for massive artificial intelligence models. However, hardware neural network reports still mainly focus on shallow networks (2 to 5 layers). Implementing deep neural networks in hardware is challenging due to the layer-by-layer structure, resulting in long training times, signal interference, and low accuracy due to gradient explosion/vanishing. Here, we utilize negative ultraviolet photoconductive light-emitting memristors with intrinsic parallelism and hardware-software co-design to achieve electrical information's optical cross-layer transmission. We propose a hybrid ultra-deep photoelectric neural network and an ultra-deep super-resolution reconstruction neural network using light-emitting memristors and cross-layer block, expanding the networks to 54 and 135 layers, respectively. Further, two networks enable transfer learning, approaching or surpassing software-designed networks in multi-dataset recognition and high-resolution restoration tasks. These proposed strategies show great potential for high-precision multifunctional hardware neural networks and edge artificial intelligence.

20 µm Micro-LEDs Mass Transfer via Laser-Induced In Situ Nanoparticles Resonance Enhancement.

Ultrafast laser is expected as a promising strategy for micro-LEDs (µ-LEDs) transfer due to its inherent property of suppressing thermal effects. However, its ultrahigh peak power and the unclear transfer mechanism make its transfer quality and efficiency unsatisfactory. Here, the study reports the high-precision mass transfer of 20 µm fine-pitch µ-LEDs via in situ nanoparticles (NPs) resonance enhancement in burst mode ultraviolet picosecond laser irradiation. This technique suppresses the thermal melting effect and rapid cooling behavior of plasma by temporal modulation of the burst mode, generating NPs-induced resonance enhancement that accurately and controllable drives a single unit up to tens of thousands of µ-LEDs. The transfer of large µ-LED arrays with more than 180 000 chips is also demonstrated, showing a transfer yield close to 99.9%, a transfer speed of 700 pcs s-1, and a transfer error of <±1.2 µm. The transferred µ-LEDs perform excellent optoelectronic properties and enable reliable device operation regardless of complex strain environments, providing a reliable strategy for preparing broader classes of 3D integrated photonics devices.

ESYT1 tethers the ER to mitochondria and is required for mitochondrial lipid and calcium homeostasis.

Mitochondria interact with the ER at structurally and functionally specialized membrane contact sites known as mitochondria-ER contact sites (MERCs). Combining proximity labelling (BioID), co-immunoprecipitation, confocal microscopy and subcellular fractionation, we found that the ER resident SMP-domain protein ESYT1 was enriched at MERCs, where it forms a complex with the outer mitochondrial membrane protein SYNJ2BP. BioID analyses using ER-targeted, outer mitochondrial membrane-targeted, and MERC-targeted baits, confirmed the presence of this complex at MERCs and the specificity of the interaction. Deletion of ESYT1 or SYNJ2BP reduced the number and length of MERCs. Loss of the ESYT1-SYNJ2BP complex impaired ER to mitochondria calcium flux and provoked a significant alteration of the mitochondrial lipidome, most prominently a reduction of cardiolipins and phosphatidylethanolamines. Both phenotypes were rescued by reexpression of WT ESYT1 and an artificial mitochondria-ER tether. Together, these results reveal a novel function for ESYT1 in mitochondrial and cellular homeostasis through its role in the regulation of MERCs.

Systematic analysis of alternative exon-dependent interactome remodeling reveals multitasking functions of gene regulatory factors.

Alternative splicing significantly expands biological complexity, particularly in the vertebrate nervous system. Increasing evidence indicates that developmental and tissue-dependent alternative exons often control protein-protein interactions; yet, only a minor fraction of these events have been characterized. Using affinity purification-mass spectrometry (AP-MS), we show that approximately 60% of analyzed neural-differential exons in proteins previously implicated in transcriptional regulation result in the gain or loss of interaction partners, which in some cases form unexpected links with coupled processes. Notably, a neural exon in Chtop regulates its interaction with the Prmt1 methyltransferase and DExD-Box helicases Ddx39b/a, affecting its methylation and activity in promoting RNA export. Additionally, a neural exon in Sap30bp affects interactions with RNA processing factors, modulating a critical function of Sap30bp in promoting the splicing of <100 nt "mini-introns" that control nuclear RNA levels. AP-MS is thus a powerful approach for elucidating the multifaceted functions of proteins imparted by context-dependent alternative exons.

Pervasive mislocalization of pathogenic coding variants underlying human disorders.

Widespread sequencing has yielded thousands of missense variants predicted or confirmed as disease-causing. This creates a new bottleneck: determining the functional impact of each variant - largely a painstaking, customized process undertaken one or a few genes or variants at a time. Here, we established a high-throughput imaging platform to assay the impact of coding variation on protein localization, evaluating 3,547 missense variants of over 1,000 genes and phenotypes. We discovered that mislocalization is a common consequence of coding variation, affecting about one-sixth of all pathogenic missense variants, all cellular compartments, and recessive and dominant disorders alike. Mislocalization is primarily driven by effects on protein stability and membrane insertion rather than disruptions of trafficking signals or specific interactions. Furthermore, mislocalization patterns help explain pleiotropy and disease severity and provide insights on variants of unknown significance. Our publicly available resource will likely accelerate the understanding of coding variation in human diseases.

SPA-UNet: A liver tumor segmentation network based on fused multi-scale features.

Liver tumor segmentation is a critical part in the diagnosis and treatment of liver cancer. While U-shaped convolutional neural networks (UNets) have made significant strides in medical image segmentation, challenges remain in accurately segmenting tumor boundaries and detecting small tumors, resulting in low segmentation accuracy. To improve the segmentation accuracy of liver tumors, this work proposes space pyramid attention (SPA)-UNet, a novel image segmentation network with an encoder-decoder architecture. SPA-UNet consists of four modules: (1) Spatial pyramid convolution block (SPCB), extracting multi-scale features by fusing three sets of dilated convolutions with different rates. (2) Spatial pyramid pooling block (SPPB), performing downsampling to reduce image size. (3) Upsample module, integrating dense positional and semantic information. (4) Residual attention block (RA-Block), enabling precise tumor localization. The encoder incorporates 5 SPCBs and 4 SPPBs to capture contextual information. The decoder consists of the Upsample module and RA-Block, and finally a segmentation head outputs segmented images of liver and liver tumor. Experiments using the liver tumor segmentation dataset demonstrate that SPA-UNet surpasses the traditional UNet model, achieving a 1.0 and 2.0% improvement in intersection over union indicators for liver and tumors, respectively, along with increased recall rates by 1.2 and 1.8%. These advancements provide a dependable foundation for liver cancer diagnosis and treatment.

Genome-scale mapping of DNA damage suppressors through phenotypic CRISPR-Cas9 screens.

To maintain genome integrity, cells must accurately duplicate their genome and repair DNA lesions when they occur. To uncover genes that suppress DNA damage in human cells, we undertook flow-cytometry-based CRISPR-Cas9 screens that monitored DNA damage. We identified 160 genes whose mutation caused spontaneous DNA damage, a list enriched in essential genes, highlighting the importance of genomic integrity for cellular fitness. We also identified 227 genes whose mutation caused DNA damage in replication-perturbed cells. Among the genes characterized, we discovered that deoxyribose-phosphate aldolase DERA suppresses DNA damage caused by cytarabine (Ara-C) and that GNB1L, a gene implicated in 22q11.2 syndrome, promotes biogenesis of ATR and related phosphatidylinositol 3-kinase-related kinases (PIKKs). These results implicate defective PIKK biogenesis as a cause of some phenotypes associated with 22q11.2 syndrome. The phenotypic mapping of genes that suppress DNA damage therefore provides a rich resource to probe the cellular pathways that influence genome maintenance.

A central chaperone-like role for 14-3-3 proteins in human cells.

14-3-3 proteins are highly conserved regulatory proteins that interact with hundreds of structurally diverse clients and act as central hubs of signaling networks. However, how 14-3-3 paralogs differ in specificity and how they regulate client protein function are not known for most clients. Here, we map the interactomes of all human 14-3-3 paralogs and systematically characterize the effect of disrupting these interactions on client localization. The loss of 14-3-3 binding leads to the coalescence of a large fraction of clients into discrete foci in a client-specific manner, suggesting a central chaperone-like function for 14-3-3 proteins. Congruently, the engraftment of 14-3-3 binding motifs to nonclients can suppress their aggregation or phase separation. Finally, we show that 14-3-3s negatively regulate the localization of the RNA-binding protein SAMD4A to cytoplasmic granules and inhibit its activity as a translational repressor. Our work suggests that 14-3-3s have a more prominent role as chaperone-like molecules than previously thought.

Ruthenium-Catalyzed Regioselective Hydrohalogenation of Alkynes Mediated by Trimethylsilyl Triflate.

Here we describe a ruthenium-catalyzed regioselective hydrohalogenation reaction of alkynes under mild conditions. Commercially simple halogen sources such as KI, ZnBr2, and ZnCl2 were employed to achieve this transformation. Alkynes derived from bioactive molecules such as l-(-)-borneol, l-menthol, and estrone were also suitable for the transformation, demonstrating the potential synthetic value of this new reaction in organic synthesis.

Approximately 30 nm Nanogroove Formation on Single Crystalline Silicon Surface under Pulsed Nanosecond Laser Irradiation.

Nanogrooves with a minimum feature size down to 30 nm (λ/26) can be formed directly on silicon surface by irradiation from two orthogonal polarized 1064 nm/10 ns fiber laser beams. The creation of such small nanogrooves is attributed to surface thermal stress during resolidification and supercooling with the double laser beams' irradiation. By varying the pulse number and laser fluence, the feature size of narrow grooves on silicon surface can be tuned. The experimental results and numerical calculation of surface thermal behaviors indicated that the high repetition rate of the nanosecond laser leads to the incubation effect and different silicon optical and thermal properties during laser irradiation. Resolution on this scale should be attractive in nanolithography, particularly considering that this method is available in far field and in ambient air.

Chemical genomics with pyrvinium identifies C1orf115 as a regulator of drug efflux.

Pyrvinium is a quinoline-derived cyanine dye and an approved anti-helminthic drug reported to inhibit WNT signaling and have anti-proliferative effects in various cancer cell lines. To further understand the mechanism by which pyrvinium is cytotoxic, we conducted a pooled genome-wide CRISPR loss-of-function screen in the human HAP1 cell model. The top drug-gene sensitizer interactions implicated the malate-aspartate and glycerol-3-phosphate shuttles as mediators of cytotoxicity to mitochondrial complex I inhibition including pyrvinium. By contrast, perturbation of the poorly characterized gene C1orf115/RDD1 resulted in strong resistance to the cytotoxic effects of pyrvinium through dysregulation of the major drug efflux pump ABCB1/MDR1. Interestingly, C1orf115/RDD1 was found to physically associate with ABCB1/MDR1 through proximity-labeling experiments and perturbation of C1orf115 led to mis-localization of ABCB1/MDR1. Our results are consistent with a model whereby C1orf115 modulates drug efflux through regulation of the major drug exporter ABCB1/MDR1.

Systematic mapping of nuclear domain-associated transcripts reveals speckles and lamina as hubs of functionally distinct retained introns.

The nucleus is highly compartmentalized through the formation of distinct classes of membraneless domains. However, the composition and function of many of these structures are not well understood. Using APEX2-mediated proximity labeling and RNA sequencing, we surveyed human transcripts associated with nuclear speckles, several additional domains, and the lamina. Remarkably, speckles and lamina are associated with distinct classes of retained introns enriched in genes that function in RNA processing, translation, and the cell cycle, among other processes. In contrast to the lamina-proximal introns, retained introns associated with speckles are relatively short, GC-rich, and enriched for functional sites of RNA-binding proteins that are concentrated in these domains. They are also highly differentially regulated across diverse cellular contexts, including the cell cycle. Thus, our study provides a resource of nuclear domain-associated transcripts and further reveals speckles and lamina as hubs of distinct populations of retained introns linked to gene regulation and cell cycle progression.

Identification and functional characterization of transcriptional activators in human cells.

Transcription is orchestrated by thousands of transcription factors (TFs) and chromatin-associated proteins, but how these are causally connected to transcriptional activation is poorly understood. Here, we conduct an unbiased proteome-scale screen to systematically uncover human proteins that activate transcription in a natural chromatin context. By combining interaction proteomics and chemical inhibitors, we delineate the preference of these transcriptional activators for specific co-activators, highlighting how even closely related TFs can function via distinct cofactors. We also identify potent transactivation domains among the hits and use AlphaFold2 to predict and experimentally validate interaction interfaces of two activation domains with BRD4. Finally, we show that many novel activators are partners in fusion events in tumors and functionally characterize a myofibroma-associated fusion between SRF and C3orf62, a potent p300-dependent activator. Our work provides a functional catalog of potent transactivators in the human proteome and a platform for discovering transcriptional regulators at genome scale.

Leveraging machine learning essentiality predictions and chemogenomic interactions to identify antifungal targets.

Fungal pathogens pose a global threat to human health, with Candida albicans among the leading killers. Systematic analysis of essential genes provides a powerful strategy to discover potential antifungal targets. Here, we build a machine learning model to generate genome-wide gene essentiality predictions for C. albicans and expand the largest functional genomics resource in this pathogen (the GRACE collection) by 866 genes. Using this model and chemogenomic analyses, we define the function of three uncharacterized essential genes with roles in kinetochore function, mitochondrial integrity, and translation, and identify the glutaminyl-tRNA synthetase Gln4 as the target of N-pyrimidinyl-ß-thiophenylacrylamide (NP-BTA), an antifungal compound.

Comprehensive interactome profiling of the human Hsp70 network highlights functional differentiation of J domains.

Hsp70s comprise a deeply conserved chaperone family that has a central role in maintaining protein homeostasis. In humans, Hsp70 client specificity is provided by 49 different co-factors known as J domain proteins (JDPs). However, the cellular function and client specificity of JDPs have largely remained elusive. We have combined affinity purification-mass spectrometry (AP-MS) and proximity-dependent biotinylation (BioID) to characterize the interactome of all human JDPs and Hsp70s. The resulting network suggests specific functions for many uncharacterized JDPs, and we establish a role of conserved JDPs DNAJC9 and DNAJC27 in histone chaperoning and ciliogenesis, respectively. Unexpectedly, we find that the J domain of DNAJC27 but not of other JDPs can fully replace the function of endogenous DNAJC27, suggesting a previously unappreciated role for J domains themselves in JDP specificity. More broadly, our work expands the role of the Hsp70-regulated proteostasis network and provides a platform for further discovery of JDP-dependent functions.

TAZ-CAMTA1 and YAP-TFE3 alter the TAZ/YAP transcriptome by recruiting the ATAC histone acetyltransferase complex.

Epithelioid hemangioendothelioma (EHE) is a vascular sarcoma that metastasizes early in its clinical course and lacks an effective medical therapy. The TAZ-CAMTA1 and YAP-TFE3 fusion proteins are chimeric transcription factors and initiating oncogenic drivers of EHE. A combined proteomic/genetic screen in human cell lines identified YEATS2 and ZZZ3, components of the Ada2a-containing histone acetyltransferase (ATAC) complex, as key interactors of both fusion proteins despite the dissimilarity of the C terminal fusion partners CAMTA1 and TFE3. Integrative next-generation sequencing approaches in human and murine cell lines showed that the fusion proteins drive a unique transcriptome by simultaneously hyperactivating a TEAD-based transcriptional program and modulating the chromatin environment via interaction with the ATAC complex. Interaction of the ATAC complex with both fusion proteins indicates that it is a key oncogenic driver and unifying enzymatic therapeutic target for this sarcoma. This study presents an approach to mechanistically dissect how chimeric transcription factors drive the formation of human cancers.

The proliferation of human cells is tightly regulated to ensure that enough cells are made to build and repair organs and tissues, while at the same time stopping cells from dividing uncontrollably and damaging the body. To get the right balance, cells rely on physical and chemical cues from their environment that trigger the biochemical signals that regulate two proteins called TAZ and YAP. These proteins control gene activity by regulating the rate at which genes are copied to produce proteins. If this process becomes dysregulated, cells can grow uncontrollably, causing cancer. In cancer cells, it is common to find TAZ and YAP fused to other proteins. In epithelioid hemangioendothelioma, a rare cancer that grows in the blood vessels, cancerous growth can be driven by a version of TAZ fused to the protein CAMTA1, or a version of YAP fused to the protein TFE3. While the role of TAZ and YAP in promoting gene activity is known, it is unclear how CAMTA1 and TFE3 contribute to cell growth becoming dysregulated. Merritt, Garcia et al. studied sarcoma cell lines to show that these two fusion proteins, TAZ-CAMTA1 and YAP-TFE3, change the pattern of gene activity seen in the cells compared to TAZ or YAP alone. An analysis of molecules that interact with the two fusion proteins identified a complex called ATAC as the cause of these changes. This complex adds chemical markers to DNA-packaging proteins, which control which genes are available for activation. The fusion proteins combine the ability of TAZ and YAP to control gene activity and the ability of CAMTA1 and TFE3 to make DNA more accessible, allowing the fusion proteins to drive uncontrolled cancerous growth. Similar TAZ and YAP fusion proteins have been found in other cancers, which can activate genes and potentially alter DNA packaging. Targeting drug development efforts at the proteins that complex with TAZ and YAP fusion proteins may lead to new therapies.

A High-Density Human Mitochondrial Proximity Interaction Network.

We used BioID, a proximity-dependent biotinylation assay with 100 mitochondrial baits from all mitochondrial sub-compartments, to create a high-resolution human mitochondrial proximity interaction network. We identified 1,465 proteins, producing 15,626 unique high-confidence proximity interactions. Of these, 528 proteins were previously annotated as mitochondrial, nearly half of the mitochondrial proteome defined by Mitocarta 2.0. Bait-bait analysis showed a clear separation of mitochondrial compartments, and correlation analysis among preys across all baits allowed us to identify functional clusters involved in diverse mitochondrial functions and to assign uncharacterized proteins to specific modules. We demonstrate that this analysis can assign isoforms of the same mitochondrial protein to different mitochondrial sub-compartments and show that some proteins may have multiple cellular locations. Outer membrane baits showed specific proximity interactions with cytosolic proteins and proteins in other organellar membranes, suggesting specialization of proteins responsible for contact site formation between mitochondria and individual organelles.

Systematic mapping of genetic interactions for de novo fatty acid synthesis identifies C12orf49 as a regulator of lipid metabolism.

The de novo synthesis of fatty acids has emerged as a therapeutic target for various diseases, including cancer. Because cancer cells are intrinsically buffered to combat metabolic stress, it is important to understand how cells may adapt to the loss of de novo fatty acid biosynthesis. Here, we use pooled genome-wide CRISPR screens to systematically map genetic interactions (GIs) in human HAP1 cells carrying a loss-of-function mutation in fatty acid synthase (FASN), whose product catalyses the formation of long-chain fatty acids. FASN-mutant cells show a strong dependence on lipid uptake that is reflected in negative GIs with genes involved in the LDL receptor pathway, vesicle trafficking and protein glycosylation. Further support for these functional relationships is derived from additional GI screens in query cell lines deficient in other genes involved in lipid metabolism, including LDLR, SREBF1, SREBF2 and ACACA. Our GI profiles also identify a potential role for the previously uncharacterized gene C12orf49 (which we call LUR1) in regulation of exogenous lipid uptake through modulation of SREBF2 signalling in response to lipid starvation. Overall, our data highlight the genetic determinants underlying the cellular adaptation associated with loss of de novo fatty acid synthesis and demonstrate the power of systematic GI mapping for uncovering metabolic buffering mechanisms in human cells.

Realization of ∼10 nm Features on Semiconductor Surfaces via Femtosecond Laser Direct Patterning in Far Field and in Ambient Air.

Direct fabrication of ∼10 nm features by optical means in far field and in ambient air on semiconductor surfaces is significant for next-generation advances nanomanufacturing. We report here a new method that enables the direct formation of 12 nm (λ/66) features on silicon surfaces. It is processed in far field and in ambient air via the irradiation of orthogonally polarized double femtosecond laser beams. The coupling of orthogonally polarized double femtosecond laser beams and the incubation effect due to multiple femtosecond laser pulses irradiation under high repetition rate enable the 12 nm nanostructures creation parallel to the scanning direction, regardless of scanning path.