Key to vaccination against infectious diseases is a formulation that facilitates appropriate kinetics to induce protective immunity. Using continuous liquid interface production (CLIP) 3D printing, faceted microneedle (MN) arrays were generated. Although purified antigens are significantly safer, they are less immunogenic, which necessitates the addition of immunostimulatory adjuvants. Antigens and adjuvants need to be delivered appropriately to innate immune cells. Increasing evidence suggests vaccine kinetics and delivery location are crucial to increasing protective immunity while minimizing adverse effects. Controlled-release exposure to antigens and lengthened exposure of antigens in draining lymph nodes by nanoparticle delivery greatly boosted humoral immunity. Alternatives to bolus delivery of vaccines include using intradermal routes, as human skin is rich in immune cells. Using biodegradable materials, MNs can be fabricated that are self-administered onto skin, eliminating the need for trained medical personnel and hypodermic needles—a source of dangerous medical waste. MNs can be stored dry, reducing the need for reconstitution reducing or removing the need for a cold supply chain. MN production unfortunately is not trivial: (1) desirable needle geometries are challenging to make by microfabrication, (2) repeated molding often blunts MNs, and (3) MN mechanical properties also must match the vaccine formulation, which is challenging. Using the novel CLIP 3D-printing technology, the authors rapidly printed faceted MN arrays to increase the MN surface area for delivering the volume of vaccine cargo. CLIP 3D printing yields high precision and consistency, generating 700 µm polyethylene glycol in 10 × 10 arrays on 10 mm × 10 mm patches. The incorporation of horizontal grooves on the MNs gave rise to a 36% increase in model drug (ovalbumin; OVA) loading. The MN patches were coated by soaking in a vaccine formulation containing OVA and CpG (5’—C—phosphate—G—3’) oligonucleotide (for immunostimulation) for a 10-second period. Each patch contained 17 µg OVA and 2.5 µg CpG. By adjusting the MN design or cargo concentration, the loading amount(s) can be readily tuned with >80% OVA and CpG load released in 2 hrs. To evaluate MN performance in vivo, fluorescently labeled OVA and CpG were monitored. Within 2 h, 80% of the CpG load was delivered as compared with 70% OVA. MN-delivered CpG showed longer persistence (up to 72 h) as compared with injected CpG (only 24 h). OVA cargo delivery characteristics were similar between both MN and injection delivery. MN delivery of CpG and OVA also accumulated to a greater degree in the draining lymph nodes (dLNs) and was more efficient in recruiting immune cells 72 h after administration. Higher numbers of antigen-presenting cells were observed, suggesting that MNs were efficient in generating an adaptive immune response. C57BL/6 mice were immunized with a double immunization 23 d apart with 16.5 µg OVA and 2.5 µg CpG for each dose. Compared with injection delivery, MNs induced 20-fold higher OVA-specific immunoglobulin G (IgG) following the first dose and 50-fold following the boost. MN delivery also promoted Th1-biased immunity, increasing total germinal center B cells in the dLNs twofold as compared with alternative routes of vaccine administration. MN delivery also achieved dose sparing, maintaining similar immune responses even with significantly reduced OVA and CpG, suggesting it is more cost-effective as an administrative route. The duration of the MN vaccine antibody response was also longer, with a response detectable at day 196, whereas other routes returned to baseline by day 119. In addition, T-cell immune responses were also significantly boosted with the MN delivery formulation. This study is a clear demonstration of the benefits of using CLIP 3D printing technology to contribute toward developing more efficient vaccination strategies to assist global immunization and public health care. (Caudill, C.; et al. Proc. Natl. Acad. Sci. U. S. A. 2021, 118, e2102595118) Cultured meat is an extremely attractive technology with benefits in ethics, economics, the environment, and even public health. Although plant-based meat analogs have already been released commercially, cultured meat may recapitulate real meat in areas such as flavor, ration of tissue constituents (e.g., muscle, adipose ratio), texture, and so forth. Although bovine cell fiber–based hamburger has been previously demonstrated, the production of realistic meat at scale and cost-effectively is challenging. Because of its scalability and the controllability of structure and composition, 3D cell printing is a promising shaping method. Of the current methods of 3D printing, supporting bath–assisted 3D printing (SBP) can be manipulated readily under high shear forces (due to its thixotropic nature) but returns to its originally high viscosity, maintaining its printed form. SBP has even been shown to fabricate tissue of complex shapes. Steak meat consists of skeletal muscle fascicles with a diameter ranging from 900 to 2300 µm depending on factors such as age and anatomy, among others. Thus, the composition of muscle, adipose tissue, and blood capillaries can vary to a great extent. Herein, the authors demonstrate a process involving (1) the collection of edible bovine satellite cells (bSCs) and bovine adipose-derived stem cells (bADSCs) before expanding their numbers through culture; (2) the development of tendon-gel–integrated bioprinting (TIP) to fabricate muscle cell fibers and the subsequent assembly of muscle, adipose, blood capillaries, and so on; and (3) the assembly of cell fibers to mimic the structure of actual steak meat. From the masseter muscle of Japanese black cows, bSCs were extracted with CD31–, CD45–, CD56+, and CD29+ cells isolated by fluorescence-activated cell sorting. These were differentiation induced by switching to 2% horse serum, which induces muscle cell differentiation. To assess bADSC adipogenic differentiation, 3D culture with collagen microfibers and fibrin gel was then performed. To mature bovine adipocytes, the transforming growth factor type I receptor activin-like kinase 5 inhibitor (ALK5i) was added to the culture media. To induce endothelial cells, horse serum increased CD31 expression in bADSCs by 15-fold compared with fetal bovine serum. The authors then printed bioink containing bSCs, fibrinogen, and Matrigel into a bath containing gelatin or gellan gum and thrombin. In these gels, high levels of viability were observed 9 d after printing, as evidenced by live/dead viability staining. To prevent the collapse of printed cell fibers, these were anchored onto silicone rubber. Thus, the SBP was modified to include components that can be anchored by the printed cell fiber. The cultured meat product was printed onto polydimethylsiloxane wells. On day 3, the printed cell fiber maintained its fibrous shape, keeping the connection with the tendon gels. A high alignment of cells was further observed, with major histocompatibility complex expression upregulated by >1000-fold on day 3 compared with naive bSCs, and the sarcomere structures showed highly matured muscle fibers. When compared with commercial beef cuts, TIP-derived muscle fibers had a DNA concentration that was 6× smaller, suggesting that the bSC seeding density and extracellular matrix (ECM) concentration needs to be optimized. The compressive moduli were also within one order of kPa. To demonstrate the construction of cultured steak, the authors then took a cross-sectional image of Wagyu with sarcomeric α-actinin and laminin stainings. They tried to produce a cultured steak of 5 mm × 10 mm × 5 mm and arranged the components of muscle, fat, blood capillary cell fibers, and so on. The food cross-linking enzyme transglutaminase was then added to accelerate assembly for 2 d at 4 °C. Finally, a cross-sectional image was obtained to demonstrate the analogousness to Wagyu, implying the TIP-based method could generate large pieces of steak meat. According to the authors, this was the first work to demonstrate the creation of Wagyu-like meat of a significant size (in centimeter dimensions) from three components (muscle, fat, blood vessels). Further work is still required to evaluate the taste, texture, and other organoleptic traits of the meat analogue. (Kang, D.; et al. Nat. Commun. 2021, 12, 5059) Stem cell (cell-replacement) therapies for neurologic disorders (such as central nervous system [CNS] injuries) have gained significant interest in recent years. Despite their potential, their clinical use is impeded by a poor survival rate, inefficient integration of neural plasticity, and uncontrollable differentiation of implanted cells due to the inhibitory and inflammatory milieu of the injury/disease sites. One approach involves using scaffold materials to facilitate favorable microenvironments for stem cell/cell implantation. However, outstanding issues limit the implementation in the clinic, including mechanical mismatches, poorly matched biodegradation properties, and immune reactions from polymer biomaterial scaffolding. Injectable hydrogels partially address the above issues. Self-assembling peptide hydrogels are immunomodulatory delivering drugs in a programmable and biocompatible manner. Injectable hydrogels, however, lack critical attributes and niches, including cell-cell interactions. Thus, scaffold approaches remain unsuitable for CNS applications. Three-dimensional cell spheroids are also very attractive; these enhance survival and differentiation, stimulating secretion of neurotrophic factors, allowing high densities of cells, potentially improving therapeutic outcomes from stem cell implantation. This technology still leaves room for improvement, because the cell-signaling cues are generated spontaneously. This uncontrolled differentiation, along with the lack of cell-ECM interactions and inhomogeneous access to aggregated cells inside the spheroid, gives rise to undesired gliogenesis and inefficient neurogenesis. The development of a “necrotic core” due to insufficient nutrients and growth factors within spheroids similarly plague these technologies. Thus, a combined approach involving stem cell spheroids, materials, and controlled drug release is potentially very attractive, as it may provide the full suite of cell-cell, cell-matrix, and therapeutic drugs released at the disease/injury site to provide the greatest synergistic benefit. To generate “SMART” neural stem cells (NSCs), a solution of 1 million cells/mL NSCs were mixed with neural ECM-laminin until cell aggregates were generated in 10 to 15 min. MNO2 nanosheets were further added to assist in binding laminin. Compared with conventional neurospheres, SMART neurospheres displayed hierarchical structures with NSCs interfacing with nanostructured 3D matrices. Importantly, these assembled an order of magnitude faster than conventional spheroid formation. Importantly, MNO2 demonstrated good biocompatibility within the work range of cell concentrations. The SMART neurospheres also expressed favorable cell-cell interactions, including higher expression of focal adhesion kinase (FAK) compared with control spheroids. This suggested reduced cell-cell interactions evidenced by increased Notch expression. Axonal growth was also further increased by 6.9-fold growth. Treating NSCs with an FAK inhibitor before NSC aggregation resulted in loosely assembled aggregates compared with the SMART neurospheres. Using the SMART neurosphere platform, DAPT, a small-molecule γ-secretase inhibitor, is adsorbed onto MnO2 nanosheets. DAPT represses Notch signaling, causing a decrease in gliogenesis and an increase in neuronal differentiation. Rhodamine B (RhB) was incorporated in the MNO2 and used to monitor DAPT in live cells. The authors suggest that the controlled release of DAPT from the SMART neurospheres can generate neuronal differentiation. As a nondegrading control experiment, graphene oxide nanosheets were compared with the MNO2 nanosheets. After 3 d of drug release, only the MNO2 group degraded, releasing the drug, as observed from RhB fluorescence. A relatively homogeneous drug diffusion profile was observed over 3 d, suggesting that the MNO2 nanosheets could facilitate drug delivery in potential disease/injury sites. Tracking the released MN2+ ions was another method to monitor SMART spheroid degradation. By doing so, the authors observed that 75% of the nanosheet was degraded by 6 d. Within the SMART neurospheres, the released DAPT led to a concentration-dependent increase in neuronal gene expression (e.g., Tuj1), whereas astrocyte genes (glial fibrillary acidic protein) were decreased. Immunostaining further showed a 3.95-fold enhancement of Tuj1 as well as increased axonal growth. The addition of basic fibroblast growth factor onto the nanosheet surfaces further enhanced cell survival. In mouse models, a hemisection was performed at thoracic levels T8-T10 and implanted with DAPT-containing SMART neurospheres (SMART-DAPT). Seven days after injection, higher levels of cell survival were observed in the SMART-DAPT condition. In vivo, there was an increase in neuronal differentiation efficiency, with 4× longer axon outgrowth compared with control conditions. One month after implantation, there was a 3.3-fold increase in cell counts compared with single-cell injection controls as well as a higher rate of neuronal differentiation (55%). Glial scar formation after injury was also reduced by 26% compared with the single-cell condition and an increase by 3.6-fold of NeuN+ cells that represent endogenous neurons in the spinal cord’s ventral horn. In addition to molecular markers, faster recovery and improved functional recovery were observed from Basso Mouse Scale scoring. By 7 d, the MNO2 nanosheets degraded, with the dark-colored material disappearing. MN2+ ions in the urine were also found to fully disappear at 28 d. Inspection of major organs such as the liver and kidneys showed no differences from the single-cell injection group. Taken together, the SMART neurosphere platform demonstrated good incorporation of neuronal-promoting drugs, maintaining implanted cell viability that led to comparative improvement in the mouse subject with negligible system toxicity. Not limiting themselves to CNS conditions, the authors envisage the use of different cell candidates, drugs, growth factors, and drugs to address a wide range of diseases and disorders. (Rathnam, C.; et al. Sci. Adv. 2021, 7, eabj2281) Although vaccines for COVID-19 are generally effective, their safety and efficacy in pregnant women have not been well studied. As a result, vaccination guidelines for pregnant women have been inconsistent, ranging from contraindicated to recommended. Although real-world studies suggest no safety issues among pregnant women, the efficacy of COVID-19 vaccines in pregnant women remains unclear. Pregnancy is also known to cause alteration in the immune system, causing decreases in CD4 and CD8 lymphocytes as well as inflammatory cytokines. Because of differences in pregnant women’s immunity, it is highly likely that the immune response from RNA vaccines will be different compared with that of the general population. The newness of RNA vaccines makes them interesting to investigate, because different immune function may affect the efficacy of the vaccines. From a cohort of Israeli pregnant women, 28,227 were eligible. A total of 10,861 vaccinated women were matched to unvaccinated women. During a median follow-up of 77 d, 131 infections were recorded in the vaccinated women, while 235 infections occurred in the control group. Promisingly, the vaccines were 96% effective against documented infections 7 to 56 d after the second vaccine dose. This suggests that RNA vaccines are considered effective in pregnant women compared with the general population. This also suggests that general population efficacy may be used to estimate the efficacy of future RNA vaccines. A further benefit is that vaccinations provide protection for newborns. Some neutralizing and binding antibodies were even found in the breast milk of vaccinated individuals. These findings make persuasive arguments for inoculating pregnant women against COVID-19. Although the current results are very promising, further studies can better identify the dynamics of vaccine effectiveness during pregnancy and the relationship between vaccination timing and infant protection. (Dagan, N.; et al. Nat. Med. 2021, doi:10.1038/s41591-021-01490-8) Despite the success of the RNA vaccines for COVID-19, clinical trials mostly excluded immunocompromised individuals, including patients on immunosuppressive strategies to control inflammatory conditions, patients with primary immunodeficiencies, recipients of organ transplants, and patients on chemotherapy. These underserved patient populations are of particular concern because the mortality rate is 10% to 30% higher than that of the general population. Immunocompromised patients have a diminished immune response to SARS-CoV-2 infections as well as mRNA vaccinations, which vary according to the cause. While those with autoimmune conditions have reduced humoral responses, organ transplant recipients mount poor antibody responses after the first dose, which subsequently rise on the second. Thus, more work needs to be performed to study and protect this population, including providing additional immunizations. Patients with a known solid tumor malignancy on active immunosuppressive cancer therapy were vaccinated with the COVID-19 BNT162b2 vaccine and compared with healthy controls. CD19+ B cells were reduced by comparison, whereas CD13+ myeloid cells were increased in the cancer cohort. In both cohorts, S2 antibodies increased after the first and second vaccinations, although these were comparatively lower in the cancer cohort. Moreover, antibodies of the receptor-binding domain (RBD) of the spike protein in the cancer cohort were also diminished at the second dosage and several days later. RBD antibody titers later showed that the cancer cohort was 11-fold lower compared with the healthy cohort. Some individuals even failed to generate any antibody response above the detection limit. Neutralizing antibody titer, which correlates well with protection from infection, was threefold higher after the first vaccine dose in the control cohort and then sixfold higher after the second dose. When detecting IFNγ+ T cells, only the second vaccine dose gave rise to a significant increase in numbers for the cancer cohort. Thereafter, an intervention trial was performed for the cancer cohort to boost their immunity with a third dose. RBD-specific antibodies and virus-neutralizing antibodies were boosted, while spike-specific T cells remained similar. Some participants with no detectable neutralizing antibodies had increases in these levels after the third dose. Interestingly, unlike the healthy cohort, there was a lack of correlation between RBD-specific memory B-cell frequencies and boosted neutralizing antibody titers. These results show that a third vaccine dose would likely boost neutralizing antibody levels, although T-cell activity may not benefit. Quantitative antibody tests could be used to identify individuals from this heterogeneous cancer cohort who would receive the greatest benefit from a third vaccine dose. (Shroff, R.; et al. Nat. Med. 2021, doi:10.1038/s41591-021-01542-z) The mRNA-1273 (Moderna) vaccine is a lipid nanoparticle-encapsulated mRNA derived from the S protein of the original Wuhan-Hu-1 isolate. It achieved 94% efficacy against symptomatic COVID-19 disease in >30,000 participants. Subsequently, several variants of concern have emerged, and there have been a few reports on reduced efficacy against the Beta and Delta variants. A previous study even showed that the neutralizing antibodies dropped between 2.1- and 8.4-fold against the Delta, Gamma, and Beta variants. This suggests the possibility of increasing breakthrough infections and waning vaccine efficacy. This study was meant to explore the safety and immunogenicity of a single booster dose and different combinations encoding the S protein from the Beta variant. With regard to the safety profile, local and systemic adverse reactions (ARs) were similar among the booster groups. The frequency of more severe symptoms (grade 3 ARs) ranged from 10% to 15%, with the most common local AR as injection site pain. Fatigue, headache, and joint and muscle pains were also common systemic ARs. Fifteen percent reported fever, with no serious ARs reported. Wild-type D614G and Beta variant neutralization was measured in samples before the booster dose and at 15 and 29 d after. Most collected samples neutralized wild-type D614G virus, whereas the neutralization titer for Beta was low or nondetectable before the booster dose. The different combinations of boosters improved the geometric mean titer (GMT) between 9.2-fold and 46.4-fold for wild-type D614G on day 29. GMTs against the Beta variant were 33.7- to 61.6-fold higher compared with day 1. A month after the primary vaccination series, wild-type D614G antibody GMT ranged from 1210 to 2213 using the pseudovirus neutralization (PsVN) assay. The assay further demonstrated that at 6 mo after the primary vaccination series, levels of neutralizing antibodies decreased compared with peak titers against wild-type D614G. Neutralizing antibody levels against the Beta and Gamma variants were undetectable for 30% to 44% of the samples. Furthermore, sera from a random subset of participants were used to assess neutralization against the Delta variant and dropped by 33- to 40-fold. Approximately half the samples even fell below detection levels. Following the booster shots, sera were collected 2 wk after and compared with samples taken after the primary vaccination series. After the boosters, wild-type D614G neutralization was between 1.7-fold and 4.4-fold higher. The boosters also increased neutralization against the variants to levels that were like those of the D614G wild-type benchmarks. Of the boosters, the multivalent mRNA-1273.211 achieved the greatest GMT against the variants. Using the multivalent mRNA-1273.351 and mRNA-1273.211 seemed to induce greater neutralizing antibody titers against the Beta variant than regular mRNA-1273 did. In certain cases, the variant neutralizing GMT following the multivalent booster was higher compared with peak wild-type D614G GMT after the primary vaccination series, suggesting it increases the breadth of coverage against variants. Although the results seem promising, it is preferred not to draw any conclusions at this stage. Yet, there are a number of limitations relating to this analysis: the treatment groups were not randomized; sample sizes were not large enough; participants were either White and non-Hispanic or Latino, which is not sufficiently broad to generalize among different races and ethnicities; sex, mean body mass index, and age were not equivalent among the booster groups; and the PsVN assay, although reliable, was still considered only “research grade.” While promising at the early stage, the absence of a correlate of protection means it cannot be definitively determined whether the neutralization titers from the multivalent boosters could protect against the SARS-CoV2 variants. This strategy of developing multivalent booster strategies could be employed in the future to vaccinate against emerging variants through the development of variant-specific vaccines. (Choi, A.; et al., Nat. Med. 2021. doi:10.1038/s41591-021-01527-y) After 22 mo from its discovery, SARS-CoV-2 continues its destructive grip on the world. Immunization remains a key player in halting this. Coronaviruses remain a dangerous foe to humankind, with two previous ones demonstrating pathogenic potential and a few more strains designated “pre-emergent” SARS-like viruses. Identifying broadly effective vaccines against dangerous coronaviruses remains a high priority. Coronavirus (CoV) spike proteins offer multiple immunogenic sites with potential for vaccine development including the S2 subunit, the receptor binding domain (RBD), and the N terminal binding domain (NTD). Although NTD-specific antibodies show promise in protecting against SARS-CoV-2 challenges, it is not clear whether vaccine-raised neutralizing antibodies can be sufficiently protective. To evaluate this prospect, lipid particles containing mRNA of RBD, NTD, and S2 units from different coronaviruses with pathogenic potential were inoculated in aged mice. The authors proceeded to design chimeric spikes containing NTD, RBD, and S2 epitopes to generate broad protective immune responses. Following confirmation of chimeric spike expression in human embryonic kidney cells, aged mice were inoculated with these sequences. Using enzyme-linked immunosorbent assay (ELISA) against a diverse range of CoV spike proteins, these chimeric spikes elicited higher-magnitude and broader responses compared with their monovalent SARS-CoV-2 counterparts. These vaccine candidates were further evaluated against live CoVs with pathogenic potential. Whereas monovalent vaccines for SARS-CoV-2 mounted a robust response against the same virus, it fared poorly against the other CoVs (18- to 300-fold lower neutralizing antibodies compared with SARS-CoV-2). On the other hand, the chimeric S vaccines (group 1) maintained antibody levels at relatively high levels against non-SARS-CoV-2 CoVs of interest. Upon challenge with classic (2003) SARS, mice were protected from weight loss as well as lower and upper airway virus replication. The monovalent SARS-CoV-2 vaccine protected against SARS-CoV-2 but not from the severe symptoms induced by classic SARS. This suggests the present SARS-CoV-2 vaccines would not protect against future pathogenic CoVs. Further analysis showed that chimeric vaccines with NTD were necessary for broad CoV protection. In NTD deletion mutant viruses such as the B.1.351 variant, chimeric vaccines also demonstrated sufficient protection in aged mice. To assess acute lung injury in mice, lung tissue sections were scored for diffuse alveolar damage. The chimeric spike vaccines provided complete protection following classic SARS challenge, while monovalent SARS-CoV-2 failed to prevent lung inflammation as in control group mice. Chimeric spike vaccines reduced eosinophilic infiltrates compared with monovalent vaccines. Similarly, proinflammatory cytokinesis and chemokines such as interleukin (IL)–6, chemokine ligand 2 (CCL2), IL-1a, granulocyte colony-stimulating factor, and CCL4 levels were at baseline levels in chimeric spike vaccines, whereas monovalent vaccines expressed them above baseline levels. This study suggests that combining distinct antigenic sequences produced viable vaccines that demonstrated the interchangeability and function plasticity of the CoV spike. Cross-protection against multiple pathogenic CoVs suggests that universal vaccines are likely. Future studies will seek to identify other sequences that generate protective, broad T-cell responses, prevent the emergence of threatening viruses, and even boost immune protection in SARS-CoV2 vaccinated individuals or convalescent patients. (Martinez, D.; et al. Science 2021, 596, 471–472) Almost 2 y into COVID-19, the situation still remains severe in many countries, with significant threats to global public health and also to economic and social burdens. The rapid and accurate diagnosis of infection plays an important role in preventing virus transmission. Diagnosis involves detecting either viral RNA or antigen antibodies. Antibody testing is important in several ways, including validating RNA methods, neutralizing viruses, vaccine development and evaluation, and seroepidemiology to assess community immunity. ELISA and colloidal gold lateral flow immunoassays (LFIAs) are the most common detection modalities for antibodies. ELISAs are accurate but technically complicated, whereas LFIAs have ease of use but are relatively insensitive and prone to misdiagnosis. Although SARS-CoV-2 IgG can be detected in saliva, it is several orders of magnitude lower concentration than serum antibodies. Organic electrochemical transistors (OECTs) are high-performance transducers and amplifiers that convert biological into electrical signals. These are highly sensitive, low cost, easily fabricated and are amenable for high-throughput, multiplexed biomarker detection. OECT devices have three electrodes (gate, source, drain), an organic semiconductor channel, and electrolyte medium connecting the gate and channel. It has also detected various biomolecules such as nucleic acids, proteins, and metabolites. To detect SARS-CoV-2 IgG, an OECT biosensor controlled by portable meter and mobile phone was developed. The SARS-CoV-2 spike protein was immobilized on the gate electrode while the IgG antibody was bound to it, giving rise to an electrical response. This configuration yields a detection limit of 1 fM in aqueous and 10 fM in serum and saliva. When compared with human serum IgG and pure bovine serum albumin, no voltage shift was observed, unlike that of SARS-CoV-2 after a 10-min incubation. Further optimization suggests pH conditions of 5 to 7.2, while a 1 µM spike protein was found to be optimal. Voltage pulses were also applied to drive the positively charged IgG to the gate, enhancing antigen and antibody binding efficiency. In saliva and serum samples, 10 fM IgG was detected following a 5-min incubation time to achieve a stable signal. Operating the testing process remotely with Bluetooth facilitates fast, rapid, point-of-care COVID-19 antibody detection for public health requirements. This versatile workflow is expected to be applied to many other diseases by antibody detection. (Liu, H.; et al., Sci. Adv. 2021, 7, eabg8387. doi:10.1126/sciadv.abg8387) Due to the COVID-19 pandemic, many countries have restricted travel, negatively affecting most tourist economies. Countries have adopted a few border-screening protocols that can be classified into four types: (1) unrestricted travel from “white-list” countries, (2) negative PCR test result from “gray-listed” countries, (3) all from “red-listed” countries to quarantine, and (4) forbidding travel from “black-listed” countries. Most nations used a combination of all four strategies. Although these color designations were based on population-level metrics, the quality of these data can be highly flawed because of underreporting, population biases, and reporting delays. This motivated the design and implementation of Eva, an algorithmic, real-time, reinforcement learning system to identify asymptomatic travelers while packaging real-time information for policy makers. The Eva system includes the following:1Passenger locator form to be completed 24 h prior to arrival2Estimating the prevalence among travelers by performing Lasso regression from high-dimensional statistics, an empirical Bayes method to estimate each type’s prevalence daily3Allocating scarce testing resources by balancing between the number of asymptomatic travelers identified versus traveler types (i.e., from their given demographics), which is insufficient information to gauge prevalence4Gray listing countries based on human policy maker input5Closing the loop by obtaining results from step 3 and updating prevalence estimates in step 2 As compared with the Eva system, random surveillance was estimated to identify only 54.1% of infected travelers during the tourist season. This suggests that to achieve the same effectiveness, random testing would have required 85% more tests at each point of entry. This outperformance by the Eva system continued in October, when arrival rates dropped, increasing the relative performance to 73.4% likely because of the changed relative scarcity of testing resources. This suggests it is a system that is advantageous when resources are scarce. Eva’s performance improved even more in the second half of peak season, when infection rates increased and testing resources became even scarcer. The widely used method of using population-level epidemiological metrics is found to be inefficient and largely ineffective in accurately predicting the prevalence of COVID-19 among arriving travelers. This suggests that travel protocols used by the European Union based on epidemiological metrics is concerning, as it is not an effective tool for identifying potential travelers to be tested. The use of Eva was further estimated to prevent an additional 6.7% of infected travelers from entering the country during the peak and off-peak seasons. A few lessons could be distilled from this nationwide exercise: design algorithms around data minimization, prioritize interpretability, and finally design the algorithm for flexibility by making it modular to allow for updating to facilitate changes in circumstances. (Bastami, H;. et al. Nature 2021. doi:10.1038/s41586-021-04014-z) Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.