Charting a landmark-driven path forward for population genetics and ancient DNA research in Africa.

Population history-focused DNA and ancient DNA (aDNA) research in Africa has dramatically increased in the past decade, enabling increasingly fine-scale investigations into the continent's past. However, while international interest in human genomics research in Africa grows, major structural barriers limit the ability of African scholars to lead and engage in such research and impede local communities from partnering with researchers and benefitting from research outcomes. Because conversations about research on African people and their past are often held outside Africa and exclude African voices, an important step for African DNA and aDNA research is moving these conversations to the continent. In May 2023 we held the DNAirobi workshop in Nairobi, Kenya and here we synthesize what emerged most prominently in our discussions. We propose an ideal vision for population history-focused DNA and aDNA research in Africa in ten years' time and acknowledge that to realize this future, we need to chart a path connecting a series of "landmarks" that represent points of consensus in our discussions. These include effective communication across multiple audiences, reframed relationships and capacity building, and action toward structural changes that support science and beyond. We concluded there is no single path to creating an equitable and self-sustaining research ecosystem, but rather many possible routes linking these landmarks. Here we share our diverse perspectives as geneticists, anthropologists, archaeologists, museum curators, and educators to articulate challenges and opportunities for African DNA and aDNA research and share an initial map toward a more inclusive and equitable future.

Real-Time Machine Learning Alerts to Prevent Escalation of Care: A Nonrandomized Clustered Pragmatic Clinical Trial.

OBJECTIVES: Machine learning algorithms can outperform older methods in predicting clinical deterioration, but rigorous prospective data on their real-world efficacy are limited. We hypothesized that real-time machine learning generated alerts sent directly to front-line providers would reduce escalations. DESIGN: Single-center prospective pragmatic nonrandomized clustered clinical trial. SETTING: Academic tertiary care medical center. PATIENTS: Adult patients admitted to four medical-surgical units. Assignment to intervention or control arms was determined by initial unit admission. INTERVENTIONS: Real-time alerts stratified according to predicted likelihood of deterioration sent either to the primary team or directly to the rapid response team (RRT). Clinical care and interventions were at the providers' discretion. For the control units, alerts were generated but not sent, and standard RRT activation criteria were used. MEASUREMENTS AND MAIN RESULTS: The primary outcome was the rate of escalation per 1000 patient bed days. Secondary outcomes included the frequency of orders for fluids, medications, and diagnostic tests, and combined in-hospital and 30-day mortality. Propensity score modeling with stabilized inverse probability of treatment weight (IPTW) was used to account for differences between groups. Data from 2740 patients enrolled between July 2019 and March 2020 were analyzed (1488 intervention, 1252 control). Average age was 66.3 years and 1428 participants (52%) were female. The rate of escalation was 12.3 vs. 11.3 per 1000 patient bed days (difference, 1.0; 95% CI, -2.8 to 4.7) and IPTW adjusted incidence rate ratio 1.43 (95% CI, 1.16-1.78; p < 0.001). Patients in the intervention group were more likely to receive cardiovascular medication orders (16.1% vs. 11.3%; 4.7%; 95% CI, 2.1-7.4%) and IPTW adjusted relative risk (RR) (1.74; 95% CI, 1.39-2.18; p < 0.001). Combined in-hospital and 30-day-mortality was lower in the intervention group (7% vs. 9.3%; -2.4%; 95% CI, -4.5% to -0.2%) and IPTW adjusted RR (0.76; 95% CI, 0.58-0.99; p = 0.045). CONCLUSIONS: Real-time machine learning alerts do not reduce the rate of escalation but may reduce mortality.

Malnutrition risk assessment using a machine learning-based screening tool: A multicentre retrospective cohort.

BACKGROUND: Malnutrition is associated with increased morbidity, mortality, and healthcare costs. Early detection is important for timely intervention. This paper assesses the ability of a machine learning screening tool (MUST-Plus) implemented in registered dietitian (RD) workflow to identify malnourished patients early in the hospital stay and to improve the diagnosis and documentation rate of malnutrition. METHODS: This retrospective cohort study was conducted in a large, urban health system in New York City comprising six hospitals serving a diverse patient population. The study included all patients aged ≥ 18 years, who were not admitted for COVID-19 and had a length of stay of ≤ 30 days. RESULTS: Of the 7736 hospitalisations that met the inclusion criteria, 1947 (25.2%) were identified as being malnourished by MUST-Plus-assisted RD evaluations. The lag between admission and diagnosis improved with MUST-Plus implementation. The usability of the tool output by RDs exceeded 90%, showing good acceptance by users. When compared pre-/post-implementation, the rate of both diagnoses and documentation of malnutrition showed improvement. CONCLUSION: MUST-Plus, a machine learning-based screening tool, shows great promise as a malnutrition screening tool for hospitalised patients when used in conjunction with adequate RD staffing and training about the tool. It performed well across multiple measures and settings. Other health systems can use their electronic health record data to develop, test and implement similar machine learning-based processes to improve malnutrition screening and facilitate timely intervention.

Accurate inference of population history in the presence of background selection.

All published methods for learning about demographic history make the simplifying assumption that the genome evolves neutrally, and do not seek to account for the effects of natural selection on patterns of variation. This is a major concern, as ample work has demonstrated the pervasive effects of natural selection and in particular background selection (BGS) on patterns of genetic variation in diverse species. Simulations and theoretical work have shown that methods to infer changes in effective population size over time (Ne(t)) become increasingly inaccurate as the strength of linked selection increases. Here, we introduce an extension to the Pairwise Sequentially Markovian Coalescent (PSMC) algorithm, PSMC+, which explicitly co-models demographic history and natural selection. We benchmark our method using forward-in-time simulations with BGS and find that our approach improves the accuracy of effective population size inference. Leveraging a high resolution map of BGS in humans, we infer considerable changes in the magnitude of inferred effective population size relative to previous reports. Finally, we separately infer Ne(t) on the X chromosome and on the autosomes in diverse great apes without making a correction for selection, and find that the inferred ratio fluctuates substantially through time in a way that differs across species, showing that uncorrected selection may be an important driver of signals of genetic difference on the X chromosome and autosomes.

The Allen Ancient DNA Resource (AADR) a curated compendium of ancient human genomes.

More than two hundred papers have reported genome-wide data from ancient humans. While the raw data for the vast majority are fully publicly available testifying to the commitment of the paleogenomics community to open data, formats for both raw data and meta-data differ. There is thus a need for uniform curation and a centralized, version-controlled compendium that researchers can download, analyze, and reference. Since 2019, we have been maintaining the Allen Ancient DNA Resource (AADR), which aims to provide an up-to-date, curated version of the world's published ancient human DNA data, represented at more than a million single nucleotide polymorphisms (SNPs) at which almost all ancient individuals have been assayed. The AADR has gone through six public releases at the time of writing and review of this manuscript, and crossed the threshold of >10,000 individuals with published genome-wide ancient DNA data at the end of 2022. This note is intended as a citable descriptor of the AADR.

Accurate detection of identity-by-descent segments in human ancient DNA.

Long DNA segments shared between two individuals, known as identity-by-descent (IBD), reveal recent genealogical connections. Here we introduce ancIBD, a method for identifying IBD segments in ancient human DNA (aDNA) using a hidden Markov model and imputed genotype probabilities. We demonstrate that ancIBD accurately identifies IBD segments >8 cM for aDNA data with an average depth of >0.25× for whole-genome sequencing or >1× for 1240k single nucleotide polymorphism capture data. Applying ancIBD to 4,248 ancient Eurasian individuals, we identify relatives up to the sixth degree and genealogical connections between archaeological groups. Notably, we reveal long IBD sharing between Corded Ware and Yamnaya groups, indicating that the Yamnaya herders of the Pontic-Caspian Steppe and the Steppe-related ancestry in various European Corded Ware groups share substantial co-ancestry within only a few hundred years. These results show that detecting IBD segments can generate powerful insights into the growing aDNA record, both on a small scale relevant to life stories and on a large scale relevant to major cultural-historical events.

Development and Validation of a Deep Learning Classifier Using Chest Radiographs to Predict Extubation Success in Patients Undergoing Invasive Mechanical Ventilation.

The decision to extubate patients on invasive mechanical ventilation is critical; however, clinician performance in identifying patients to liberate from the ventilator is poor. Machine Learning-based predictors using tabular data have been developed; however, these fail to capture the wide spectrum of data available. Here, we develop and validate a deep learning-based model using routinely collected chest X-rays to predict the outcome of attempted extubation. We included 2288 serial patients admitted to the Medical ICU at an urban academic medical center, who underwent invasive mechanical ventilation, with at least one intubated CXR, and a documented extubation attempt. The last CXR before extubation for each patient was taken and split 79/21 for training/testing sets, then transfer learning with k-fold cross-validation was used on a pre-trained ResNet50 deep learning architecture. The top three models were ensembled to form a final classifier. The Grad-CAM technique was used to visualize image regions driving predictions. The model achieved an AUC of 0.66, AUPRC of 0.94, sensitivity of 0.62, and specificity of 0.60. The model performance was improved compared to the Rapid Shallow Breathing Index (AUC 0.61) and the only identified previous study in this domain (AUC 0.55), but significant room for improvement and experimentation remains.

Derivation, External Validation and Clinical Implications of a deep learning approach for intracranial pressure estimation using non-cranial waveform measurements.

Importance: Increased intracranial pressure (ICP) is associated with adverse neurological outcomes, but needs invasive monitoring. Objective: Development and validation of an AI approach for detecting increased ICP (aICP) using only non-invasive extracranial physiological waveform data. Design: Retrospective diagnostic study of AI-assisted detection of increased ICP. We developed an AI model using exclusively extracranial waveforms, externally validated it and assessed associations with clinical outcomes. Setting: MIMIC-III Waveform Database (2000-2013), a database derived from patients admitted to an ICU in an academic Boston hospital, was used for development of the aICP model, and to report association with neurologic outcomes. Data from Mount Sinai Hospital (2020-2022) in New York City was used for external validation. Participants: Patients were included if they were older than 18 years, and were monitored with electrocardiograms, arterial blood pressure, respiratory impedance plethysmography and pulse oximetry. Patients who additionally had intracranial pressure monitoring were used for development (N=157) and external validation (N=56). Patients without intracranial monitors were used for association with outcomes (N=1694). Exposures: Extracranial waveforms including electrocardiogram, arterial blood pressure, plethysmography and SpO2. Main Outcomes and Measures: Intracranial pressure > 15 mmHg. Measures were Area under receiver operating characteristic curves (AUROCs), sensitivity, specificity, and accuracy at threshold of 0.5. We calculated odds ratios and p-values for phenotype association. Results: The AUROC was 0.91 (95% CI, 0.90-0.91) on testing and 0.80 (95% CI, 0.80-0.80) on external validation. aICP had accuracy, sensitivity, and specificity of 73.8% (95% CI, 72.0%-75.6%), 99.5% (95% CI 99.3%-99.6%), and 76.9% (95% CI, 74.0-79.8%) on external validation. A ten-percentile increment was associated with stroke (OR=2.12; 95% CI, 1.27-3.13), brain malignancy (OR=1.68; 95% CI, 1.09-2.60), subdural hemorrhage (OR=1.66; 95% CI, 1.07-2.57), intracerebral hemorrhage (OR=1.18; 95% CI, 1.07-1.32), and procedures like percutaneous brain biopsy (OR=1.58; 95% CI, 1.15-2.18) and craniotomy (OR = 1.43; 95% CI, 1.12-1.84; P < 0.05 for all). Conclusions and Relevance: aICP provides accurate, non-invasive estimation of increased ICP, and is associated with neurological outcomes and neurosurgical procedures in patients without intracranial monitoring.

Medieval DNA from Soqotra points to Eurasian origins of an isolated population at the crossroads of Africa and Arabia.

Soqotra, an island situated at the mouth of the Gulf of Aden in the northwest Indian Ocean between Africa and Arabia, is home to ~60,000 people subsisting through fishing and semi-nomadic pastoralism who speak a Modern South Arabian language. Most of what is known about Soqotri history derives from writings of foreign travellers who provided little detail about local people, and the geographic origins and genetic affinities of early Soqotri people has not yet been investigated directly. Here we report genome-wide data from 39 individuals who lived between ~650 and 1750 CE at six locations across the island and document strong genetic connections between Soqotra and the similarly isolated Hadramawt region of coastal South Arabia that likely reflects a source for the peopling of Soqotra. Medieval Soqotri can be modelled as deriving ~86% of their ancestry from a population such as that found in the Hadramawt today, with the remaining ~14% best proxied by an Iranian-related source with up to 2% ancestry from the Indian sub-continent, possibly reflecting genetic exchanges that occurred along with archaeologically documented trade from these regions. In contrast to all other genotyped populations of the Arabian Peninsula, genome-level analysis of the medieval Soqotri is consistent with no sub-Saharan African admixture dating to the Holocene. The deep ancestry of people from medieval Soqotra and the Hadramawt is also unique in deriving less from early Holocene Levantine farmers and more from groups such as Late Pleistocene hunter-gatherers from the Levant (Natufians) than other mainland Arabians. This attests to migrations by early farmers having less impact in southernmost Arabia and Soqotra and provides compelling evidence that there has not been complete population replacement between the Pleistocene and Holocene throughout the Arabian Peninsula. Medieval Soqotra harboured a small population that showed qualitatively different marriage practices from modern Soqotri, with first-cousin unions occurring significantly less frequently than today.

Long shared haplotypes identify the Southern Urals as a primary source for the 10th century Hungarians.

During the Hungarian Conquest in the 10th century CE, the early medieval Magyars, a group of mounted warriors from Eastern Europe, settled in the Carpathian Basin. They likely introduced the Hungarian language to this new settlement area, during an event documented by both written sources and archaeological evidence. Previous archaeogenetic research identified the newcomers as migrants from the Eurasian steppe. However, genome-wide ancient DNA from putative source populations has not been available to test alternative theories of their precise source. We generated genome-wide ancient DNA data for 131 individuals from candidate archaeological contexts in the Circum-Uralic region in present-day Russia. Our results tightly link the Magyars to people of the Early Medieval Karayakupovo archaeological horizon on both the European and Asian sides of the southern Urals. Our analyes show that ancestors of the people of the Karayakupovo archaeological horizon were established in the Southern Urals by the Iron Age and that their descendants persisted locally in the Volga-Kama region until at least the 14th century.