Mediaboss Marketing

GEO dataset integration and immune landscape of CD

We constructed a combined dataset covering 279 CD samples and 224 control samples from mucosa after the removal of batch effects (Fig.2A,B). A broadly uncoordinated immune response is an indispensable hallmark of CD. With the aim of revealing the immune landscape, we scored the immune cell infiltration of CD patients and controls via the ssGSEA method. As illustrated in Fig.2C, the infiltration of 20 immune cells in the CD group and control group was significantly different, among which only the scores of T helper 17 (Th17) cells were lower in CD tissues than in control tissues. We then performed a correlation analysis of distinct immune cells, as shown in Fig.2D. Interestingly, Th17 cells, CD56bright natural killer (NK) cells, CD56dim NK cells and monocytes showed inverse correlations with almost all other immune cells, whereas the other immune cells were generally positively correlated with one another, which deserves special attention.

GEO dataset combination and immune landscape of CD. (A) PCA between datasets before removal of batch effects. (B) PCA between integrated datasets after removal of batch effects. (C) Infiltration levels of 28 immune cell subtypes in CD samples and controls. The blue bars represent controls, and the red bars represent CD samples. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. (D) Pearson correlation analysis of distinct immune cells. The purple squares represent positive correlations, and the orange squares represent inverse correlations. GEO Gene Expression Omnibus, CD Crohns disease, PCA principal component analysis.

A total of 1265 DEGs, consisting of 592 upregulated and 673 downregulated genes, were identified through differential expression analysis (Fig.3A). A list of possible PRGs was produced from previous research (Supplementary file 1: Table S1). Subsequently, we intersected the 1265 DEGs with 930 PRGs via a Venn diagram; thus, 130 DE-PRGs were identified (Fig.3B), which were further grouped in a heatmap (Fig.3C). The overall expression of these DE-PRGs in the CD group and control group is shown in Supplementary file 3: Fig. S1. We could conclude that the vast majority of DE-PRGs were expressed at higher levels in CD tissues than in control tissues.

Identification of DE-PRGs. (A) Volcano map of the DEGs with the cutoff threshold set at |log2 (fold change)|>1 and adj. p<0.05. The blue dots represent downregulated DEGs, the red dots represent upregulated DEGs, and the gray dots represent genes with no significant difference. (B) Venn diagram of DEGs and PRGs. Pink circle represents DEGs, blue circle represents PRGs, and their overlapping area represents DE-PRGs. (C) Clustered heatmap of the top 40 DE-PRGs. Each row represents one of the top 40 DE-PRGs, and each column represents one sample, either CD or normal. DE-PRGs differentially expressed PANoptosis-related genes, DEGs differentially expressed genes, PRGs PANoptosis-related genes, CD Crohns disease.

We then examined the latent functions and signaling pathways of the DE-PRGs. GO analysis revealed that these DE-PRGs were predominantly involved in regulation of apoptotic signaling pathway, leukocyte cellcell adhesion, regulation of inflammatory response (biological process); membrane raft, membrane microdomain, focal adhesion (cellular component); ubiquitin-like protein ligase binding, ubiquitin protein ligase binding, and phosphatase binding (molecular function) (Supplementary file 4: Fig. S2A). Additionally, DE-PRGs were notably enriched in apoptosis, proteoglycans in cancer, NOD-like receptor signaling pathway, among others, according to the KEGG results (Supplementary file 4: Fig. S2B). Moreover, a PPI network analysis of the DE-PRGs was performed and a complex network of the DE-PRGs was constructed (Supplementary file 5: Fig. S3).

To screen the hub DE-PRGs, we first capitalized on three algorithms, LASSO, SVM and RF, and discovered 20, 34 and 33 potential hub DE-PRGs, respectively (Fig.4AE). Afterward, 10 hub DE-PRGs were identified through the intersection of the machine learning results, namely CD44, cell death inducing DFFA like effector c (CIDEC), N-myc downstream regulated 1 (NDRG1), nuclear mitotic apparatus protein 1 (NUMA1), proliferation and apoptosis adaptor protein 15 (PEA15), recombination activating 1 (RAG1), S100 calcium binding protein A8 (S100A8), S100 calcium binding protein A9 (S100A9), TIMP metallopeptidase inhibitor 1 (TIMP1) and X-box binding protein 1 (XBP1) (Fig.4F). Next, we probed their interactions, as shown in Fig.4G. Most hub DE-PRGs, such as CD44, PEA15, S100A8, S100A9, TIMP1 and XBP1, were closely interrelated. Moreover, NDRG1, NUMA1 and RAG1 generally presented antagonistic effects on the other hub DE-PRGs. Finally, the diagnostic value of each hub DE-PRG in predicting CD was calculated based on our combined dataset (Fig.4H). All 10 hub DE-PRGs exhibited outstanding predictive performance with area under the curve (AUC) values greater than 0.740. Notably, the AUC reached as high as 0.871 when the 10 hub DE-PRGs were combined (Fig.4H). In addition, we conducted external validation on the GSE102133 and GSE207022 datasets, respectively. The results were satisfactory, with high AUC values (Supplementary file 6: Fig. S4).

Identification of the hub DE-PRGs. (A) Cross-validations of adjusted parameter selection in the LASSO model. Each curve corresponds to one gene. (B) LASSO coefficient analysis. Vertical dashed lines are plotted at the best lambda. (C) SVM algorithm for hub gene selection. (D) Relationship between the number of random forest trees and error rates. (E) Ranking of the relative importance of genes. (F) Venn diagram showing the 10 hub DE-PRGs identified by LASSO, SVM and RF. Pink circle represents potential hub DE-PRGs identified by RF, blue circle represents potential hub DE-PRGs identified by SVM, green circle represents potential hub DE-PRGs identified by LASSO, and their overlapping area represents the final hub DE-PRGs. (G) Chord diagram showing the correlations between the hub DE-PRGs. Red represents positive correlations between different genes and green represents negative correlations between different genes. (H) ROC curves of the hub DE-PRGs in CD diagnosis. DE-PRGs differentially expressed PANoptosis-related genes, LASSO least absolute shrinkage and selection operator, RF random forest, SVM support vector machine, ROC receiver operating characteristic, AUC area under the curve, CD Crohns disease.

Spearman correlation analysis was carried out to determine the interactions between the hub DE-PRGs and immune cells (Fig.5). CD44, PEA15, S100A8, S100A9, TIMP1 and XBP1 demonstrated noteworthy positive correlations with the infiltration of an abundance of immune cells, except for certain immune cells, such as monocytes and CD56bright NK cells. In contrast, NDRG1, NUMA1, and RAG1 were negatively associated with most types of immune cells, excluding a few immune cells such as monocytes. In addition, the CIDEC fell somewhere between these two extremes.

Spearman correlation analysis of hub DE-PRGs with immune cells. The correlations between CD44 (A), CIDEC (B), NDRG1 (C), NUMA1 (D), PEA15 (E), RAG1 (F), S100A8 (G), S100A9 (H), TIMP1 (I) and XBP1 (J) gene expressions with immune cells, respectively. The size of the dots represents the strength of gene correlation with immune cells; the larger the dot, the stronger the correlation. The color of the dots represents the p-value; the greener the color, the lower the p-value. p<0.05 was considered statistically significant. DE-PRGs differentially expressed PANoptosis-related genes.

The top 30 crucial genes related to CD were extracted from the GeneCards database, and their expression levels were compared between CD samples and normal samples (Fig.6A). We could easily conclude that a majority of the CD-related genes (26 out of 30) were differentially expressed, especially COL1A1, CTLA4, IL10 and NOD2. Pearson correlation analysis was subsequently conducted to scrutinize the relationships between these CD-related genes and the hub DE-PRGs (Fig.6B). Notably, CTLA4, one of the most differentially expressed CD-related genes, was significantly associated with each hub DE-PRG. COL1A1, IL10 and NOD2 also presented varying levels of correlation with the hub DE-PRGs. Nevertheless, there were no significant correlations between the hub DE-PRGs and some CD-related genes, including CYBB, IL10RA, RET and VCP.

Expression levels of the top 30 CD-related genes and relationships between them and hub DE-PRGs. (A) Boxplot of the top 30 crucial genes in relation to CD. The blue bars represent controls, and the red bars represent CD samples. (B) Pearson correlation analysis between the top 30 CD-related genes and the 10 hub DE-PRGs. *p<0.05; **p<0.01; ***p<0.001. CD Crohns disease, DE-PRGs differentially expressed PANoptosis-related genes.

Subsequently, a genemiRNA interaction network of the 10 hub DE-PRGs consisting of 226 nodes and 338 edges was constructed (Supplementary file 7: Fig. S5 and Supplementary file 8: Table S3). Apparently, miR-124-3p, miR-34a-5p and miR-27a-3p were most strongly associated with the hub DE-PRGs in CD. After that, we generated a geneTF regulatory network of the 10 hub DE-PRGs (Supplementary file 9: Fig. S6). The 10 hub DE-PRGs were regulated by 35 total TFs. Among them, FOXC1 was found to regulate as many as 7 hub DE-PRGs and S100A8 was regulated by 13 miRNAs (Supplementary file 10: Table S4). In addition, we looked for available drugs that act on the hub DE-PRGs, and a host of drugs were involved (Supplementary file 11: Fig. S7 and Supplementary file 12: Table S5). Specifically, a total of 19 drugs interacted with XBP1, 8 of which inhibited it.

To distinguish different PANoptosis patterns in CD patients, we adopted the NMF method for unsupervised clustering on the basis of the 10 hub DE-PRGs. At k=2, the most stable and optimal PANclusters were identified (Fig.7A). There were 101 and 178 CD samples in PANcluster A and PANcluster B, respectively. The geometrical distance between the two clusters is shown in Fig.7B, validating their gene expression heterogeneity. Thereafter, a boxplot and a heatmap were generated to compare the expression levels of the hub DE-PRGs between PANcluster A and PANcluster B (Fig.7C,D). Specifically, PANcluster A was distinguished by the considerably high expression levels of CIDEC, NDRG1, NUMA1 and RAG1, while the other hub DE-PRGs, that is, CD44, PEA15, S100A8, S100A9, TIMP1 and XBP1, were expressed at higher levels in PANcluster B.

Recognition of PANclusters in CD. (A) Unsupervised clustering matrix generated using NMF method when k=2. (B) PCA plot showing the distribution of PANcluster A and PANcluster B. The red dots represent PANcluster A and the blue dots represent PANcluster B. (C) Boxplot of the expression levels of the hub DE-PRGs in PANcluster A and PANcluster B. The red bars represent PANcluster A, and the blue bars represent PANcluster B. (D) Heatmap of the expression levels of the hub DE-PRGs in PANcluster A and PANcluster B. Each row represents one hub DE-PRG, and each column represents one CD sample. PANclusters PANoptosis patterns, CD Crohns disease, NMF nonnegative matrix factorization, PCA principal component analysis, DE-PRGs differentially expressed PANoptosis-related genes.

GSVA was performed with the aim of shedding light on the functional diversity patterns of the recognized PANclusters. With regard to Hallmark pathways, increased activity of p53 pathway, androgen response and hypoxia were detected in PANcluster A, whereas mTORC1 signaling, inflammatory response, TNF- signaling via NF-B, IL-6/JAK/STAT3 signaling and epithelial mesenchymal transition were increased in PANcluster B (Supplementary file 13: Fig. S8A). In addition, results from the KEGG analysis suggested that PANcluster A had hypoactive ECMreceptor interaction and endocytosis but expressed high levels of genes associated with cytokinecytokine receptor interaction and numerous signaling pathways, including toll-like receptor signaling pathway and NOD-like receptor signaling pathway (Supplementary file 13: Fig. S8B). Concerning the Reactome-based pathways, PANcluster A showed an increase in the cell cycle pathway, while most pathways, such as cytokine signaling in immune system and extracellular matrix-related pathways, were significantly enriched in PANcluster B (Supplementary file 13: Fig. S8C).

To clarify the disparities in the immune system among the PANclusters, we compared their immune microenvironments, as shown in Fig.8A. Remarkably, the enrichment scores of 26 immune cells were much greater in PANcluster B than in PANcluster A. Consequently, CD56bright NK cells and monocytes were the only two exceptions with higher infiltration degrees in PANcluster A, the explanations behind which demand further investigation. In addition, differential gene analysis revealed 533 DEGs, including 171 upregulated and 362 downregulated genes (Fig.8B). To learn more about the biological functions and processes linked to these DEGs, GO and KEGG analyses were performed. The 533 DEGs were markedly enriched in the following terms: positive regulation of cell adhesion, leukocyte cellcell adhesion, and extracellular matrix organization (biological process); collagen-containing extracellular matrix, secretory granule membrane, and basement membrane (cellular component); and extracellular matrix structural constituent, glycosaminoglycan binding, and integrin binding (molecular function) (Fig.8C,D). Moreover, the 533 DEGs were principally involved in many pathways, such as cell adhesion molecules, ECMreceptor interaction and PI3K-Akt signaling pathway (Fig.8E).

Characterization of different PANclusters. (A) Infiltration levels of 28 immune cell subtypes in PANclusters A and B. The red bars represent PANcluster A, and the blue bars represent PANcluster B. (B) Volcano map of DEGs between PANclusters A and B. The blue dots represent downregulated DEGs, the red dots represent upregulated DEGs, and the gray dots represent genes with no significant difference. (C,D) Enriched items in GO analysis based on the DEGs between PANclusters A and B. (E) Enriched items in KEGG analysis based on the DEGs between PANclusters A and B. Node color indicates gene expression level; quadrilateral color indicates z-score. PANclusters PANoptosis patterns, DEGs differentially expressed genes, BP biological process, CC cellular component, MF molecular function, GO Gene Ontology, KEGG Kyoto Encyclopedia of Genes and Genomes.

CD and control samples were acquired from 10 patients who were diagnosed with CD, and their demographic and clinical information is presented in Table 1. qRT-PCR was subsequently conducted to determine the relative expression levels of the 10 hub DE-PRGs (Fig.9A). As expected, the levels of CD44, PEA15, S100A8, S100A9, TIMP1 and XBP1 increased in CD samples compared with those in control samples; while the opposite trend was observed for NDRG1. Moreover, there was no significant difference in the mRNA expression levels of CIDEC, NUMA1 or RAG1. Furthermore, we established classic TNBS and DSS mouse models of CD and collected colon tissues to analyze the expression levels of the hub DE-PRGs in murine colon tissues from the TNBS, DSS and control groups (Fig.9B,C). Generally, the results of the TNBS model were in line with expectations. Specifically, in TNBS-induced colitis, Cd44, Numa1, S100a8, S100a9, Timp1 and Xbp1 were more highly expressed, while Cidec and Rag1 were less expressed. In addition, the levels of Ndrg1 and Pea15a did not significantly differ between the TNBS group and the control group. Consistent with previous work, in the DSS mouse model, the expression levels of Cd44, S100a8, S100a9 and Timp1 were greater in the mice with colitis; while the expression level of Ndrg1 was lower in the mice with colitis. In addition, no significant difference in the expression levels of Cidec, Pea15a or Xbp1 was detected. Unexpectedly, the expression levels of Numa1 and Rag1 in the DSS group were different from those in the CD and TNBS colitis groups.

qRT-PCR validation of the hub DE-PRGs in CD patients (A), TNBS-induced colitis model (B) and DSS-induced colitis model (C). The blue dots represent the normal/control tissues, and the red dots represent the diseased tissues. qRT-PCR quantitative real-time PCR, DE-PRGs differentially expressed PANoptosis-related genes, CD Crohns disease, TNBS 2,4,6-trinitrobenzene sulfonic acid, DSS dextran sodium sulfate, GAPDH glyceraldehyde-3-phosphate dehydrogenase.

Continued here:
Characterization of PANoptosis-related genes in Crohn's disease by integrated bioinformatics, machine learning and ... - Nature.com

Tags:

Understanding how proteins fold is key to unlocking how cells function and how their structures are built. When proteins fold incorrectly, it can cause cells to malfunction, leading to diseases like Alzheimer's and Parkinson's. Delving into the complexities of protein folding not only enriches our fundamental understanding of biological processes but also paves the way for innovative therapies, and novel drug discoveries.

The 2024 Boden Research Conference aims to bring together a multidisciplinary community of experts to deepen our understanding of protein folding dynamics and propel therapeutic innovation. The conference will assess current knowledge, identify crucial gaps, and guide future research in protein folding and misfolding. The conference will also explore how machine learning can revolutionise the search for new treatments and better diagnosis of diseases.

Join us to discover how the smallest proteins can solve some of our biggest health challenges. Together, we can advance scientific understanding and develop new solutions to combat debilitating diseases.

Date: 25th -26th September, 2024

Time: 8:00am -17:00 pm. A conference dinner will be organised on day one (25th September, 2024) at 18:00 pm

Deadline: Deadline for conference registration is 23:59 13th September, 2024 (AEST) , and deadline for abstract submission is 23:59 15th July, 2024 (AEST).

Cost: Free

Venue:Holmes Building (A09), Science Rd, The University of Sydney, Camperdown NSW 2050, Australia

Program: TBA

Contact: Professor Ken-Tye Yong, The University of Sydney: ken.yong@sydney.edu.au; Dr Morning Liu, The University of Sydney: xiaochen.liu@sydney.edu.au

Read the original post:
Australian Academy of Science Boden Research Conference 2024: Protein Folding: Mechanisms, Health, and Machine ... - Australian Academy of Science

Tags:

In the seemingly never-ending war to protect their customers and institutions from fraud, an increasing number of financial institutions (FIs) are deploying machine learning (ML) and artificial intelligence (AI) tools to fight back.

And according to PYMNTS Intelligences Leveraging AI and ML to Thwart Scammers, a report created in collaboration withHawk, those efforts appear to be working.

The report, which is based on surveys with 200 FIs with more than $1 billion in assets under management, revealed that thoseFIs that now use ML or AI to mitigate fraud are seeing steep declines in common forms of fraud.

For instance, tech support impersonation and IRS imposter scams are two of the mostfrequently reported scams, yet FIs using ML or AI anti-fraud solutions were 17% less likely to report experiencing these leading scams than FIs relying solely on more traditional fraud prevention tools. Likewise, they were 18% less likely to report IRS imposter scams as a top concern.

They also reported lower rates of lottery, romance, utility, rentalandSocial Security scams.In fact,as the figure illustrates, FIs leveragingthe MLand AI technology reported lower incidents of nearly every common form of fraud.

The data also finds there is some room for improvement in both ML- and AI-based solutions.The tools were less successful in identifyingcharitable-donationscams. They also missed some fake debt-collection scams. This could be because these two scams are less common (and there would thus be less data for the solutions to work from).

Despite these small hurdles, FIs areapparentlyimpressed. Fifty-two percent of the FIs we surveyed plan to implement or increase their use of ML or AI fraud prevention models.In fact, FIs using AI or ML are 17% more likely to have plans in place to implement additional ML or AI solutions than their counterparts that do not use ML or AI fraud prevention solutions. In other words, many of the FIs now using the advancedtechnologies are ready to expand their ML and AItool chest.

Our study also found that by adopting ML or AI fraud-prevention models, FIs are not only stopping more bad actors from inflicting damage but also increasing the confidence their customers have that their accounts are protected. So, in turn, customer satisfaction rates are likely to increase as fraud levels decline.

For all PYMNTS AI coverage, subscribe to the daily AINewsletter.

Go here to read the rest:
52% of FIs Plan to Lean on ML and AI to Combat Fraud - PYMNTS.com

Tags:

Summary

Machine learning is sometimes considered a subset of overarching AI. But in the context of digital security, it may be better understood as a driving force, the fuel powering the engine.

The terms AI and machine learning are often used interchangeably by professionals outside the technology, managed IT and cybersecurity trades. But, truth be told, they are separate and distinct tools that can be coupled to power digital defense systems and frustrate hackers.

Artificial iIntelligence has emerged as an almost ubiquitous part of modern life. We experience its presence in everyday household robots and the familiar Alexa voice that always seems to be listening. Practical uses of AI mimic and take human behavior one step further. In cybersecurity, it can deliver 24/7 monitoring, eliminating the need for a weary flesh-and-blood guardian to stand a post.

Machine learning is sometimes considered a subset of overarching AI. But in the context of digital security, it may be better understood as a driving force, the fuel powering the engine. Using programmable algorithms, it recognizes sometimes subtle patterns. This proves useful when deployed to follow the way employees and other legitimate network users navigate systems. Although even discussions regarding AI and machine learning feel redundant, to some degree, they are a powerful one-two punch in terms of automating security decisions.

Integrating AI calls for a comprehensive understanding of mathematics, logical reasoning, cognitive sciencesand a working knowledge of business networks. The professionals who implement AI for security purposes must also possess high-level expertise and protection planning skills. Used as a problem-solving tool, AI can provide real-time alerts and take pre-programmed actions. But it cannot effectively stem the tide of bad actors without support. Enter machine learning.

In this context, machine learning emphasizes software solutions driven by data analysis. Unlike human information processing limitations, machine learning can handle massive swaths of data. What machine learning learns, for lack of a better word, translates into actionable security intel for the overarching AI umbrella.

Some people think about machine learning as a subcategory of AI, which it is. Others comprehend it in a functional way,i.e., two sides to the same coin. But for cybersecurity experts determined to deter, detectand repel threat actors, machine learning is the gasoline that powers AI engines.

Its now essential to leverage machine learning capabilities to develop a so-called intelligent computer that can defend itself, to some degree. Although the relationship between AI and machine learning is diverse and complex, an expert can integrate them into a cybersecurity posture with relative ease. Its simply a matter of repetition and the following steps.

When properly orchestrated and refined to detect user patterns and subtle anomalies, the AI-machine learning relationship helps cybersecurity professionals keep valuable and sensitive digital assets away from prying eyes and greedy digital hands.

First and foremost, its crucial to put AI and machine learning benefits in context. Studies consistently conclude that more than 80% of all cybersecurity failures are caused by human error. Using automated technologies removes many mistake-prone employees and other network users from the equation. Along with minimizing risk, these are benefits of onboarding these automated next-generation technologies.

Improved cybersecurity efficiency. According to the 2023 Global Security Operations Center Study, cybersecurity professionals spend one-third of their workday chasing downfalse positives. This waste of time negatively impacts their ability to respond to legitimate threats, leaving a business at higher than necessary risk. The strategic application of AI and machine learning can be deployed to recognize harmless anomalies and alert a CISO or vCISO only when authentic threats are present.

Increased threat hunting capabilities.Without proactive, automated security measures like MDR (managed detection and response), organizations are too often following an outdated break-and-fix model. Hackers breach systems or deposit malware, and then the IT department spends the remainder of their day, or week, trying to purge the threat and repair the damage. Cybersecurity experts have widely adopted the philosophy that the best defense is a good offense. A thoughtful AI-machine learning strategy can engage in threat hunting without ever needing a coffee break.

Cure business network vulnerabilities.Vulnerability management approaches generally employ technologies that provide proactive automation. They close cybersecurity gaps and cure inherent vulnerabilities by identifying these weaknesses and alerting human decision-makers. Unlike scheduling a routine annual risk assessment, these cutting-edge technologies deliver ongoing analytics and constant vigilance.

Resolve cybersecurity skills gap.Its something of an open secret that there are not enough trained, certified cybersecurity experts to fill corporate positions. Thats one of the reasons why industry leaders tend to outsource managed IT and cybersecurity to third-party firms. Outsourcing helps to onboard the high-level knowledge and skills required to protect valuable digital assets and sensitive information. Without enough cybersecurity experts to safeguard businesses, automation allows the resources available to companies to drill down and identify true threats. Without these advanced technologies being used to bolster network security, its likely the number of debilitating cyberattacks would grow exponentially.

The type of predictive analytics and swift decision-making capabilities this two-prong approach delivers has seemingly endless industry applications. Banking and financial sector organizations can not only use AI and machine learning to repel hackers but also ferret out fraud. Healthcare organizations have a unique opportunity to exceed Health Insurance Portability and Accountability Act (HIPAA) requirements due to the advanced personal identity record protections it affords. Companies conducting business in the global marketplace can also get a leg-up in meeting the EUs General Data Protection Regulation (GDPR) designed to further informational privacy.

Perhaps the greatest benefit organizations garner from AI and machine learning security automation is the ability to detect, respondand expel threat actors and malicious applications. Managed IT cybersecurity experts can help companies close the skills gap by integrating these and other advanced security strategies.

John Funk is a Creative Consultant at SevenAtoms. A lifelong writer and storyteller, he has a passion for tech and cybersecurity. When hes not found enjoying craft beer or playing Dungeons & Dragons, John can be often found spending time with his cats

Check out our free e-newsletters to read more great articles..

Read this article:
How Machine Learning Revolutionizes Automation Security with AI-Powered Defense - Automation.com

Tags:

Sepkoski, J. J. A factor analytic description of the phanerozoic marine fossil record. Paleobiology 7, 3653 (1981).

Article Google Scholar

Quental, T. B. & Marshall, C. R. Diversity dynamics: molecular phylogenies need the fossil record. Trends Ecol. Evol. 25, 434441 (2010).

Article PubMed Google Scholar

Ezard, T. H., Aze, T., Pearson, P. N. & Purvis, A. Interplay between changing climate and species ecology drives macroevolutionary dynamics. Science 332, 349351 (2011).

Article ADS CAS PubMed Google Scholar

Benton, M. J. Exploring macroevolution using modern and fossil data. Proc. R. Soc. B: Biol. Sci. 282, 20150569 (2015).

Article Google Scholar

Niklas, K. J. Measuring the tempo of plant death and birth. N. Phytol. 207, 254256 (2015).

Article Google Scholar

Rabosky, D. L. & Hurlbert, A. H. Species richness at continental scales is dominated by ecological limits. Am. Nat. 185, 572583 (2015).

Article PubMed Google Scholar

Harmon, L. J. & Harrison, S. Species diversity is dynamic and unbounded at local and continental scales. Am. Nat. 185, 584593 (2015).

Article PubMed Google Scholar

Sepkoski Jr, J. Phanerozoic overview of mass extinction. In Patterns and Processes in the History of Life: Report of the Dahlem Workshop on Patterns and Processes in the History of Life Berlin 1985, June 1621, 277295 (Springer, 1986).

Benton, M. J. & Emerson, B. C. How did life become so diverse? the dynamics of diversification according to the fossil record and molecular phylogenetics. Palaeontology 50, 2340 (2007).

Article Google Scholar

Alroy, J. Geographical, environmental and intrinsic biotic controls on phanerozoic marine diversification. Palaeontology 53, 12111235 (2010).

Article Google Scholar

Weber, M. G., Wagner, C. E., Best, R. J., Harmon, L. J. & Matthews, B. Evolution in a community context: on integrating ecological interactions and macroevolution. Trends Ecol. Evol. 32, 291304 (2017).

Article PubMed Google Scholar

Niklas, K. J., Tiffney, B. H. & Knoll, A. H. Patterns in vascular land plant diversification. Nature 303, 614 616 (1983).

Article Google Scholar

Foote, M., Miller, A., Raup, D. & Stanley, S.Principles of Paleontology (W. H. Freeman, 2007). https://books.google.ch/books?id=8TsDC2OOvbYC

Close, R., Benson, R., Saupe, E., Clapham, M. & Butler, R. The spatial structure of phanerozoic marine animal diversity. Science 368, 420424 (2020).

Article ADS CAS PubMed Google Scholar

Raja, N. B. et al. Colonial history and global economics distort our understanding of deep-time biodiversity. Nat. Ecol. Evol. 6, 145154 (2022).

Article PubMed Google Scholar

Smith, A. B. & McGowan, A. J. The ties linking rock and fossil records and why they are important for palaeobiodiversity studies. Geol. Soc. Lond. Spec. Publ. 358, 17 (2011).

Article ADS Google Scholar

Benson, R., Butler, R., Close, R., Saupe, E. & Rabosky, D. Biodiversity across space and time in the fossil record. Curr. Biol. 31, R1225R1236 (2021).

Article CAS PubMed Google Scholar

Smith, A. B. Largescale heterogeneity of the fossil record: implications for phanerozoic biodiversity studies. Philos. Trans. R. Soc. Lond. Ser. B: Biol. Sci. 356, 351367 (2001).

Article CAS Google Scholar

Alroy, J. Fair sampling of taxonomic richness and unbiased estimation of origination and extinction rates. Paleontol. Soc. Pap. 16, 5580 (2010).

Article Google Scholar

Chao, A. & Jost, L. Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size. Ecology 93, 25332547 (2012).

Article PubMed Google Scholar

Raup, D. Taxonomic diversity estimation using rarefaction. Paleobiology 1, 333342 (1975).

Article Google Scholar

Alroy, J. et al. Effects of sampling standardization on estimates of phanerozoic marine diversification. Proc. Natl Acad. Sci. 98, 62616266 (2001).

Article ADS CAS PubMed PubMed Central Google Scholar

Starrfelt, J. & Liow, L. H. How many dinosaur species were there? fossil bias and true richness estimated using a poisson sampling model. Philos. Trans. R. Soc. B: Biol. Sci. 371, 20150219 (2016).

Article Google Scholar

Flannery-Sutherland, J. T., Silvestro, D. & Benton, M. J. Global diversity dynamics in the fossil record are regionally heterogeneous. Nat. Commun. 13, 117 (2022).

Article Google Scholar

Chao, A. Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43, 783791 (1987).

Alroy, J. Limits to species richness in terrestrial communities. Ecol. Lett. 21, 17811789 (2018).

Article PubMed Google Scholar

Alroy, J. On four measures of taxonomic richness. Paleobiology 46, 158175 (2020).

Article Google Scholar

Close, R., Evers, S., Alroy, J. & Butler, R. How should we estimate diversity in the fossil record? testing richness estimators using sampling-standardised discovery curves. Methods Ecol. Evol. 9, 13861400 (2018).

Article Google Scholar

Close, R. et al. The apparent exponential radiation of phanerozoic land vertebrates is an artefact of spatial sampling biases. Proc. R. Soc. B 287, 20200372 (2020).

Article PubMed PubMed Central Google Scholar

Antell, G. T., Benson, R. B. & Saupe, E. E. Spatial standardization of taxon occurrence dataa call to action. Paleobiology https://doi.org/10.1017/pab.2023.36 (2024).

Dunne, E. M., Thompson, S. E., Butler, R. J., Rosindell, J. & Close, R. A. Mechanistic neutral models show that sampling biases drive the apparent explosion of early tetrapod diversity. Nat. Ecol. Evol. 7, 14801489 (2023).

Article PubMed PubMed Central Google Scholar

Hauffe, T., Pires, M. M., Quental, T. B., Wilke, T. & Silvestro, D. A quantitative framework to infer the effect of traits, diversity and environment on dispersal and extinction rates from fossils. Methods Ecol. Evol. 13, 12011213 (2022).

Article Google Scholar

Cermeo, P. et al. Post-extinction recovery of the phanerozoic oceans and biodiversity hotspots. Nature 607, 507511 (2022).

Article ADS PubMed PubMed Central Google Scholar

Hagen, O. et al. gen3sis: a general engine for eco-evolutionary simulations of the processes that shape earths biodiversity. PLoS Biol. 19, e3001340 (2021).

Article CAS PubMed PubMed Central Google Scholar

Hagen, O., Skeels, A., Onstein, R. E., Jetz, W. & Pellissier, L. Earth history events shaped the evolution of uneven biodiversity across tropical moist forests. Proc. Natl Acad. Sci. 118, e2026347118 (2021).

Article CAS PubMed PubMed Central Google Scholar

Vilhena, D. A. & Smith, A. B. Spatial bias in the marine fossil record. PLoS One 8, e74470 (2013).

Article ADS CAS PubMed PubMed Central Google Scholar

Raup, D. M. Taxonomic diversity during the phanerozoic: the increase in the number of marine species since the paleozoic may be more apparent than real. Science 177, 10651071 (1972).

Article ADS CAS PubMed Google Scholar

Raup, D. M. Species diversity in the phanerozoic: a tabulation. Paleobiology 2, 279288 (1976).

Article Google Scholar

Foote, M., Crampton, J. S., Beu, A. G. & Nelson, C. S. Aragonite bias, and lack of bias, in the fossil record: lithological, environmental, and ecological controls. Paleobiology 41, 245265 (2015).

Article Google Scholar

Silvestro, D., Salamin, N. & Schnitzler, J. Pyrate: a new program to estimate speciation and extinction rates from incomplete fossil data. Methods Ecol. Evol. 5, 11261131 (2014).

Article Google Scholar

Cantalapiedra, J. L. et al. The rise and fall of proboscidean ecological diversity. Nat. Ecol. Evol. 5, 12661272 (2021).

Article PubMed Google Scholar

Rumelhart, D. E., Hinton, G. E. & Williams, R. J. Learning representations by back-propagating errors. Nature 323, 533536 (1986).

Article ADS Google Scholar

Hochreiter, S. & Schmidhuber, J. Long short-term memory. Neural Comput. 9, 17351780 (1997).

Article CAS PubMed Google Scholar

Gers, F., Schmidhuber, J. & Cummins, F. Learning to forget: continual prediction with lstm. Neural Comput. 12, 24512471 (2000).

Article CAS PubMed Google Scholar

Gal, Y. & Ghahramani, Z. A theoretically grounded application of dropout in recurrent neural networks. Adv. Neural Inform. Process. Syst. 29, 19 (2016).

Gal, Y. & Ghahramani, Z. Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In International Conference on Machine Learning 48, 10501059 (PMLR, 2016).

Silvestro, D. & Andermann, T. Prior choice affects ability of bayesian neural networks to identify unknowns. arXiv preprint arXiv:2005.04987 (2020).

Brusatte, S. L. et al. The extinction of the dinosaurs. Biol. Rev. 90, 628642 (2015).

Article PubMed Google Scholar

Dunne, E. M., Farnsworth, A., Greene, S. E., Lunt, D. J. & Butler, R. J. Climatic drivers of latitudinal variation in late triassic tetrapod diversity. Palaeontology 64, 101117 (2021).

Article Google Scholar

De Celis, A., Narvez, I., Arcucci, A. & Ortega, F. Lagersttte effect drives notosuchian palaeodiversity (crocodyliformes, notosuchia). Historical Biol. 33, 30313040 (2021).

Article Google Scholar

Cleary, T. J., Benson, R. B., Holroyd, P. A. & Barrett, P. M. Tracing the patterns of non-marine turtle richness from the triassic to the palaeogene: from origin to global spread. Palaeontology 63, 753774 (2020).

Article Google Scholar

Silvestro, D. et al. Fossil data support a pre-Cretaceous origin of flowering plants. Nat. Ecol. Evol. 5, 449457 (2021).

Leuenberger, C. & Wegmann, D. Bayesian computation and model selection without likelihoods. Genetics 184, 243252 (2010).

Article PubMed PubMed Central Google Scholar

Marjoram, P., Molitor, J., Plagnol, V. & Tavar, S. Markov chain monte carlo without likelihoods. Proc. Natl Acad. Sci. 100, 1532415328 (2003).

Article ADS CAS PubMed PubMed Central Google Scholar

Go here to see the original:
DeepDive: estimating global biodiversity patterns through time using deep learning - Nature.com

Tags: