In the last few years, the deep learning (DL) computing paradigm has been deemed the Gold Standard in the machine learning (ML) community. Moreover, it has gradually become the most widely used computational approach in the field of ML, thus achieving outstanding results on several complex cognitive tasks, matching or even beating those provided by human performance. One of the benefits of DL is the ability to learn massive amounts of data. The DL field has grown fast in the last few years and it has been extensively used to successfully address a wide range of traditional applications. More importantly, DL has outperformed well-known ML techniques in many domains, e.g., cybersecurity, natural language processing, bioinformatics, robotics and control, and medical information processing, among many others. Despite it has been contributed several works reviewing the State-of-the-Art on DL, all of them only tackled one aspect of the DL, which leads to an overall lack of knowledge about it. Therefore, in this contribution, we propose using a more holistic approach in order to provide a more suitable starting point from which to develop a full understanding of DL. Specifically, this review attempts to provide a more comprehensive survey of the most important aspects of DL and including those enhancements recently added to the field. In particular, this paper outlines the importance of DL, presents the types of DL techniques and networks. It then presents convolutional neural networks (CNNs) which the most utilized DL network type and describes the development of CNNs architectures together with their main features, e.g., starting with the AlexNet network and closing with the High-Resolution network (HR.Net). Finally, we further present the challenges and suggested solutions to help researchers understand the existing research gaps. It is followed by a list of the major DL applications. Computational tools including FPGA, GPU, and CPU are summarized along with a description of their influence on DL. The paper ends with the evolution matrix, benchmark datasets, and summary and conclusion.

/pdf/review-of-deep-learning-concepts-cnn-architectures-hpbk62qzc2.pdf

Review of deep learning: concepts, CNN architectures, challenges, applications, future directions

Ribosomes are highly conserved ribonucleoprotein nanomachines that translate information in the genome to create the proteome in all cells. In yeast these complex particles contain four RNAs (>5400 nucleotides) and 79 different proteins. During the past 25 years, studies in yeast have led the way to understanding how these molecules are assembled into ribosomes in vivo. Assembly begins with transcription of ribosomal RNA in the nucleolus, where the RNA then undergoes complex pathways of folding, coupled with nucleotide modification, removal of spacer sequences, and binding to ribosomal proteins. More than 200 assembly factors and 76 small nucleolar RNAs transiently associate with assembling ribosomes, to enable their accurate and efficient construction. Following export of preribosomes from the nucleus to the cytoplasm, they undergo final stages of maturation before entering the pool of functioning ribosomes. Elaborate mechanisms exist to monitor the formation of correct structural and functional neighborhoods within ribosomes and to destroy preribosomes that fail to assemble properly. Studies of yeast ribosome biogenesis provide useful models for ribosomopathies, diseases in humans that result from failure to properly assemble ribosomes.

https://www.genetics.org/content/genetics/195/3/643.full.pdf

Ribosome Biogenesis in the Yeast Saccharomyces cerevisiae

pLink, software for data analysis of cross-linked proteins coupled with mass spectrometry, estimates false discovery rate and enables analysis of protein complexes without extensive purification. We have developed pLink, software for data analysis of cross-linked proteins coupled with mass-spectrometry analysis. pLink reliably estimates false discovery rate in cross-link identification and is compatible with multiple homo- or hetero-bifunctional cross-linkers. We validated the program with proteins of known structures, and we further tested it on protein complexes, crude immunoprecipitates and whole-cell lysates. We show that it is a robust tool for protein-structure and protein-protein–interaction studies.

Identification of cross-linked peptides from complex samples

rRNAs are extensively modified during their transcription and subsequent maturation in the nucleolus, nucleus and cytoplasm. RNA modifications, which are installed either by snoRNA-guided or by stand-alone enzymes, generally stabilize the structure of the ribosome. However, they also cluster at functionally important sites of the ribosome, such as the peptidyltransferase center and the decoding site, where they facilitate efficient and accurate protein synthesis. The recent identification of sites of substoichiometric 2'-O-methylation and pseudouridylation has overturned the notion that all rRNA modifications are constitutively present on ribosomes, highlighting nucleotide modifications as an important source of ribosomal heterogeneity. While the mechanisms regulating partial modification and the functions of specialized ribosomes are largely unknown, changes in the rRNA modification pattern have been observed in response to environmental changes, during development, and in disease. This suggests that rRNA modifications may contribute to the translational control of gene expression.

https://tandfonline.com/doi/pdf/10.1080/15476286.2016.1259781

Tuning the ribosome: The influence of rRNA modification on eukaryotic ribosome biogenesis and function

Abstract Characterization of the genetic landscape of Alzheimer’s disease (AD) and related dementias (ADD) provides a unique opportunity for a better understanding of the associated pathophysiological processes. We performed a two-stage genome-wide association study totaling 111,326 clinically diagnosed/‘proxy’ AD cases and 677,663 controls. We found 75 risk loci, of which 42 were new at the time of analysis. Pathway enrichment analyses confirmed the involvement of amyloid/tau pathways and highlighted microglia implication. Gene prioritization in the new loci identified 31 genes that were suggestive of new genetically associated processes, including the tumor necrosis factor alpha pathway through the linear ubiquitin chain assembly complex. We also built a new genetic risk score associated with the risk of future AD/dementia or progression from mild cognitive impairment to AD/dementia. The improvement in prediction led to a 1.6- to 1.9-fold increase in AD risk from the lowest to the highest decile, in addition to effects of age and the APOE ε4 allele. 

/pdf/new-insights-into-the-genetic-etiology-of-alzheimers-disease-2xdh9agn.pdf

New insights into the genetic etiology of Alzheimer’s disease and related dementias

Stimulated cells and cancer cells have widespread shortening of mRNA 3'-untranslated regions (3'UTRs) and switches to shorter mRNA isoforms due to usage of more proximal polyadenylation signals (PASs) in introns and last exons. U1 snRNP (U1), vertebrates' most abundant non-coding (spliceosomal) small nuclear RNA, silences proximal PASs and its inhibition with antisense morpholino oligonucleotides (U1 AMO) triggers widespread premature transcription termination and mRNA shortening. Here we show that low U1 AMO doses increase cancer cells' migration and invasion in vitro by up to 500%, whereas U1 over-expression has the opposite effect. In addition to 3'UTR length, numerous transcriptome changes that could contribute to this phenotype are observed, including alternative splicing, and mRNA expression levels of proto-oncogenes and tumor suppressors. These findings reveal an unexpected role for U1 homeostasis (available U1 relative to transcription) in oncogenic and activated cell states, and suggest U1 as a potential target for their modulation.

/pdf/u1-snrnp-regulates-cancer-cell-migration-and-invasion-in-4wdiaemkkl.pdf

U1 snRNP regulates cancer cell migration and invasion in vitro

Structural Mechanism of Substrate RNA Recruitment in H/ACA RNA-Guided Pseudouridine Synthase

Box H/ACA ribonucleoprotein particles (RNPs) mediate pseudouridine synthesis, ribosome formation, and telomere maintenance. The structure of eukaryotic H/ACA RNPs remains poorly understood. We reconstituted functional Saccharomyces cerevisiae H/ACA RNPs with recombinant proteins Cbf5, Nop10, Gar1, and Nhp2 and a two-hairpin H/ACA RNA; determined the crystal structure of a Cbf5, Nop10, and Gar1 ternary complex at 1.9 Aresolution; and analyzed the structure-function relationship of the yeast complex. Although eukaryotic H/ACA RNAs have a conserved two-hairpin structure, isolated single-hairpin RNAs are also active in guiding pseudouridylation. Nhp2, unlike its archaeal counterpart, is largely dispensable for the activity, reflecting a functional adaptation of eukaryotic H/ACA RNPs to the variable RNA structure that Nhp2 binds. The N-terminal extension of Cbf5, a hot spot for dyskeratosis congenita mutation, forms an extra structural layer on the PUA domain. Gar1 is distinguished from the assembly factor Naf1 by containing a C-terminal extension that controls substrate turnover and the Gar1-Naf1 exchange during H/ACA RNP maturation. Our results reveal significant novel features of eukaryotic H/ACA RNPs. (Keywords: pseudouridylation; H/ACA RNA-protein complex; crystal structure; dyskeratosis congenita) Supplemental material is available for this article.

/pdf/reconstitution-and-structural-analysis-of-the-yeast-box-h-38k4ksx346.pdf

Reconstitution and structural analysis of the yeast box H/ACA RNA-guided pseudouridine synthase

The higher abundance of U1 among snRNPs is explained by the identification of an additional mode of assembly: RBP U1-70K bridges pre-U1 to SMN–Gemin2–Sm, thus establishing a Gemin5-independent assembly pathway.

/pdf/a-u1-snrnp-specific-assembly-pathway-reveals-the-smn-complex-2hr7avdldd.pdf

A U1 snRNP-specific assembly pathway reveals the SMN complex as a versatile hub for RNP exchange

Shq1 is a conserved protein required for the biogenesis of eukaryotic H/ACA ribonucleoproteins (RNPs), including human telomerase. We report the structure of the Shq1-specific domain alone and in complex with H/ACA RNP proteins Cbf5, Nop10 and Gar1. The Shq1-specific domain adopts a novel helical fold and primarily contacts the PUA domain and the otherwise disordered C-terminal extension (CTE) of Cbf5. The structure shows that dyskeratosis congenita mutations found in the CTE of human Cbf5 likely interfere with Shq1 binding. However, most mutations in the PUA domain are not located at the Shq1-binding surface and also have little effect on the yeast Cbf5–Shq1 interaction. Shq1 binds Cbf5 independently of the H/ACA RNP proteins Nop10, Gar1 and Nhp2 and the assembly factor Naf1, but shares an overlapping binding surface with H/ACA RNA. Shq1 point mutations that disrupt Cbf5 interaction suppress yeast growth particularly at elevated temperatures. Our results suggest that Shq1 functions as an assembly chaperone that protects the Cbf5 protein complexes from non-specific RNA binding and aggregation before assembly of H/ACA RNA.

https://www.embopress.org/doi/epdf/10.1038/emboj.2011.427

Jingqi Duan

Papers

U1 snRNP regulates cancer cell migration and invasion in vitro

Structural Mechanism of Substrate RNA Recruitment in H/ACA RNA-Guided Pseudouridine Synthase

Reconstitution and structural analysis of the yeast box H/ACA RNA-guided pseudouridine synthase

A U1 snRNP-specific assembly pathway reveals the SMN complex as a versatile hub for RNP exchange

Structure of the Shq1–Cbf5–Nop10–Gar1 complex and implications for H/ACA RNP biogenesis and dyskeratosis congenita