SpaiNN: equivariant message passing for excited-state nonadiabatic molecular dynamics.

Excited-state molecular dynamics simulations are crucial for understanding processes like photosynthesis, vision, and radiation damage. However, the computational complexity of quantum chemical calculations restricts their scope. Machine learning offers a solution by delivering high-accuracy properties at lower computational costs. We present SpaiNN, an open-source Python software for ML-driven surface hopping nonadiabatic molecular dynamics simulations. SpaiNN combines the invariant and equivariant neural network architectures of SchNetPack with SHARC for surface hopping dynamics. Its modular design allows users to implement and adapt modules easily. We compare rotationally-invariant and equivariant representations in fitting potential energy surfaces of multiple electronic states and properties arising from the interaction of two electronic states. Simulations of the methyleneimmonium cation and various alkenes demonstrate the superior performance of equivariant SpaiNN models, improving accuracy, generalization, and efficiency in both training and inference.

Robotic scrub nurse to anticipate surgical instruments based on real-time laparoscopic video analysis.

BACKGROUND: Machine learning and robotics technologies are increasingly being used in the healthcare domain to improve the quality and efficiency of surgeries and to address challenges such as staff shortages. Robotic scrub nurses in particular offer great potential to address staff shortages by assuming nursing tasks such as the handover of surgical instruments. METHODS: We introduce a robotic scrub nurse system designed to enhance the quality of surgeries and efficiency of surgical workflows by predicting and delivering the required surgical instruments based on real-time laparoscopic video analysis. We propose a three-stage deep learning architecture consisting of a single frame-, temporal multi frame-, and informed model to anticipate surgical instruments. The anticipation model was trained on a total of 62 laparoscopic cholecystectomies. RESULTS: Here, we show that our prediction system can accurately anticipate 71.54% of the surgical instruments required during laparoscopic cholecystectomies in advance, facilitating a smoother surgical workflow and reducing the need for verbal communication. As the instruments in the left working trocar are changed less frequently and according to a standardized procedure, the prediction system works particularly well for this trocar. CONCLUSIONS: The robotic scrub nurse thus acts as a mind reader and helps to mitigate staff shortages by taking over a great share of the workload during surgeries while additionally enabling an enhanced process standardization.

Staff shortages in healthcare are an emerging problem leading to undersupply of medical experts such as scrub nurses in the operating room. The absence of these scrub nurses, who are responsible for providing surgical instruments, means that surgeries must be postponed or canceled. Robotic technologies and artificial intelligence offer great potential to address staff shortages in the operating room. We developed a robotic scrub nurse system that is able to take over the tasks of a human scrub nurse by delivering the required surgical tools. To maintain the pace of the surgery, our robotic scrub nurse system is also capable of predicting these required surgical tools in advance using artificial intelligence. The system is tested on laparoscopic cholecystectomies, a surgery, where the gallbladder is removed in a minimally invasive technique. We show that our prediction system can predict the majority of surgical instruments for this specific surgery facilitating a smoother surgical workflow and reducing the need for verbal communication. With further development, our system may help to cover the need for surgery while streamlining the surgical process through predictive support, potentially improving patient outcomes.

Towards multimodal graph neural networks for surgical instrument anticipation.

PURPOSE: Decision support systems and context-aware assistance in the operating room have emerged as the key clinical applications supporting surgeons in their daily work and are generally based on single modalities. The model- and knowledge-based integration of multimodal data as a basis for decision support systems that can dynamically adapt to the surgical workflow has not yet been established. Therefore, we propose a knowledge-enhanced method for fusing multimodal data for anticipation tasks. METHODS: We developed a holistic, multimodal graph-based approach combining imaging and non-imaging information in a knowledge graph representing the intraoperative scene of a surgery. Node and edge features of the knowledge graph are extracted from suitable data sources in the operating room using machine learning. A spatiotemporal graph neural network architecture subsequently allows for interpretation of relational and temporal patterns within the knowledge graph. We apply our approach to the downstream task of instrument anticipation while presenting a suitable modeling and evaluation strategy for this task. RESULTS: Our approach achieves an F1 score of 66.86% in terms of instrument anticipation, allowing for a seamless surgical workflow and adding a valuable impact for surgical decision support systems. A resting recall of 63.33% indicates the non-prematurity of the anticipations. CONCLUSION: This work shows how multimodal data can be combined with the topological properties of an operating room in a graph-based approach. Our multimodal graph architecture serves as a basis for context-sensitive decision support systems in laparoscopic surgery considering a comprehensive intraoperative operating scene.

A Plea for Monitoring Serum Selenium Levels in Breast Cancer Patients: Selenium Deficiency Is Rare during the First Year of Therapy, and Selenium Supplementation Is Associated with Elevated Risk of Overdosing.

(1) Background: The role of selenium in cancer biology remains poorly understood. Our aim was to study the course of selenium serum levels and the use of selenium supplements during breast cancer therapy. (2) Methods: Serum selenium levels, clinical-pathological data, selenium supplementation, and lifestyle factors were monitored quarterly over one year. (3) Results: A total of 110 non-metastatic breast cancer patients were enrolled in the prospective observational "BEGYN-1" study. At baseline, 2.9% of patients were selenium-deficient (<50 ng/mL), 1.9% were overdosed (>120 ng/mL), and 6.4% received substitution. The median selenium level was 81.5 ng/mL and ranged between 78.7 and 84.5 ng/mL within the year. A total of 25.3% of the patients received supplementation, resulting in significantly higher selenium levels (p < 0.05). A total of 8.7-28.6% of the patients using supplements were overdosed. Selenium levels strongly correlated with mushroom consumption (p = 0.003), but no association was found with therapy or clinical characteristics. (4) Conclusions: Although selenium deficiency is rare, serum selenium levels should be assessed in breast cancer patients. Mushrooms and nuts should be preferred over supplements to correct selenium deficiency. Ruling out selenium deficiency helps prevent the risk of selenosis and avoid unnecessary, costly supplementation in patients who are often financially burdened due to their disease.

Course of Vitamin D Levels in Newly Diagnosed Non-Metastatic Breast Cancer Patients over One Year with Quarterly Controls and Substitution.

(1) Background: Vitamin D levels in patients remain inadequately understood, with research yielding inconsistent findings. Breast cancer patients, particularly due to oncological therapies, face an increased risk of osteopenia, which can be exacerbated by a vitamin D deficiency. (2) Methods: The prospective observational "BEGYN-1" study assessed serum 25(OH)D levels at baseline and quarterly thereafter. Clinical, pathological, nutritional, vitamin supplementation, and lifestyle data were recorded. (3) Results: Before treatment, 68.5% of patients were vitamin D deficient (<30 ng/mL), with 4.6% experiencing severe deficiency (<10 ng/mL). The median baseline 25(OH)D levels were 24 ng/mL (range: 4.8 to 64.7 ng/mL). Throughout the study, the median vitamin D levels increased to 48 ng/mL (range: 22.0 to 76.7 ng/mL). Before diagnosis, 16.7% received vitamin D substitution, and 97.8% received vitamin D substitution throughout the year with a median weekly dose of 20,000 IU. It took at least three quarterly assessments for 95% of patients to reach the normal range. A multiple GEE analysis identified associations between 25(OH)D levels and supplementation, season, age, VLDL, magnesium levels, and endocrine therapy. (4) Conclusions: Physicians should monitor 25(OH)D levels before, during, and after oncological therapy to prevent vitamin D deficiency and to adjust substitution individually. While variables such as seasons, age, VLDL, magnesium, diet, and oncological interventions affect 25(OH)D levels, supplementation has the greatest impact.

Digital decision support for structural improvement of melanoma tumor boards: using standard cases to optimize workflow.

PURPOSE: Choosing optimal cancer treatment is challenging, and certified cancer centers must present all patients in multidisciplinary tumor boards (MDT). Our aim was to develop a decision support system (DSS) to provide treatment recommendations for apparently simple cases already at conference registration and to classify these as "standard cases". According to certification requirements, discussion of standard cases is optional and would thus allow more time for complex cases. METHODS: We created a smartphone query that simulated a tumor conference registration and requested all information needed to provide a recommendation. In total, 111 out of 705 malignant melanoma cases discussed at a skin cancer center from 2017 to 2020 were identified as potential standard cases, for which a digital twin recommendation was then generated by DSS. RESULTS: The system provided reliable advice in all 111 cases and showed 97% concordance of MDT and DSS for therapeutic recommendations, regardless of tumor stage. Discrepancies included two cases (2%) where DSS advised discussions at MDT and one case (1%) with deviating recommendation due to advanced patient age. CONCLUSIONS: Our work aimed not to replace clinical expertise but to alleviate MDT workload and enhance focus on complex cases. Overall, our DSS proved to be a suitable tool for identifying standard cases as such, providing correct treatment recommendations, and thus reducing the time burden of tumor conferences in favor for the comprehensive discussion of complex cases. The aim is to implement the DSS in routine tumor board software for further qualitative assessment of its impact on oncological care.

Diagnostic performance of additional imaging tests for staging purposes in a bicentric German series of low-risk early breast cancer patients.

PURPOSE: Low-risk early breast cancer rarely leads to the development of metastatic disease, and in these patients, additional imaging test is controversial. The aim of our study was to evaluate the conventional staging procedures in a bicentric German series of low-risk breast carcinoma patients. METHODS: Retrospective evaluation of all patients diagnosed with early, low-risk breast cancer at Saarland University Hospital and Freiburg University Hospital in 2017 was performed. Clinical patient characteristics, the number and type of additional imaging examinations, follow-up examinations, and results were evaluated. The detection rate of metastases and the rate of false-positive findings were analyzed. RESULTS: A total of 203 patients were included, with all patients received at least one additional imaging test. Initially, a total of 562 additional imaging examinations were performed: 166 chest X-rays, 169 upper abdominal ultrasounds, 199 bone scans, 27 computer tomographies (CT) chest and abdomen, and 1 CT abdomen. 6.8% of patients had abnormal findings reported, requiring 38 additional imaging examinations. One patient (0.5%) was found to have bone metastases. The rate of false-positive findings in the performed additional imaging procedures was 6.6%. CONCLUSION: Metastatic disease was detected in one of 203 patients with low-risk early breast cancer. A total of 562 examinations and additional 38 follow-up examinations were performed without detection of metastasis (this corresponds to approximately 3 examinations/patient). The rate of false-positive findings was 6.6%. The performance of additional imaging procedures for detection of distant metastases should be critically reconsidered in patients with low-risk early breast cancer.

New Twist on the Light-Switch Effect: Controlling the Fate of Excited States with pH in a 4-Hydroxy-thiazol-extended Ruthenium(II) Dppz Complex.

We present a pH-dependent study of the excited state dynamics of a novel Ru complex bearing a 4-hydroxy thiazol-substituted dppz (dipyridophenazine) ligand (RuTzOH) and its deprotonated form (RuTzO-). We combine steady-state and time-resolved absorption and emission spectroscopy with electrochemical investigations to characterize the excited state relaxation, which upon photoexcitation at 400 nm is determined by a multitude of initially populated MLCT states for both complexes. Subsequently, for RuTzOH, two long-lived excited states are populated, leading to dual emission from the complexes, a feature that vanishes upon deprotonation. Upon deprotonation, the electron density on the dppz moiety increases significantly, leading to rapid energy populating ligand-centered states and thus deactivating the initially excited MLCT states.

Design, construction, and functional characterization of a tRNA neochromosome in yeast.

Here, we report the design, construction, and characterization of a tRNA neochromosome, a designer chromosome that functions as an additional, de novo counterpart to the native complement of Saccharomyces cerevisiae. Intending to address one of the central design principles of the Sc2.0 project, the ∼190-kb tRNA neochromosome houses all 275 relocated nuclear tRNA genes. To maximize stability, the design incorporates orthogonal genetic elements from non-S. cerevisiae yeast species. Furthermore, the presence of 283 rox recombination sites enables an orthogonal tRNA SCRaMbLE system. Following construction in yeast, we obtained evidence of a potent selective force, manifesting as a spontaneous doubling in cell ploidy. Furthermore, tRNA sequencing, transcriptomics, proteomics, nucleosome mapping, replication profiling, FISH, and Hi-C were undertaken to investigate questions of tRNA neochromosome behavior and function. Its construction demonstrates the remarkable tractability of the yeast model and opens up opportunities to directly test hypotheses surrounding these essential non-coding RNAs.

Mammary gland, kidney and rumen urea and uric acid transporters of dairy cows differing in milk urea concentration.

The milk urea concentration (MUC) serves as indicator of urinary nitrogen emissions, but at comparable crude protein (CP) intake, cows with high (HMU) and low (LMU) MUC excrete equal urea amounts. We hypothesized that urea and uric acid transporters and sizes of the kidney, mammary gland, and rumen account for these phenotypes. Eighteen HMU and 18 LMU Holstein dairy cows fed a low (LP) and normal (NP) CP diet were studied. Milk, plasma and urinary urea concentrations were greater with NP feeding, while plasma and urinary urea concentrations were comparable between phenotypes. Milk and plasma uric acid concentrations were higher with LP feeding but not affected by phenotype. The milk-urine uric acid ratio was greater in HMU cows. The mRNA expressions of the ruminal urea transporter SLC14A1 and AQP10, the mammary gland and rumen AQP3, and the mammary gland uric acid transporter ABCG2 were not affected by group or diet. Renal AQP10, but not AQP3, AQP7, and SLC14A2 expressions, and the kidney weights were lower in HMU cows. These data indicate that renal size and AQP10 limit the urea transfer from blood to urine, and that MUC determines if uric acid is more released with milk or urine.

Analysis of protein secretion in Bacillus subtilis by combining a secretion stress biosensor strain with an in vivo split GFP assay.

BACKGROUND: Bacillus subtilis is one of the workhorses in industrial biotechnology and well known for its secretion potential. Efficient secretion of recombinant proteins still requires extensive optimization campaigns and screening with activity-based methods. However, not every protein can be detected by activity-based screening. We therefore developed a combined online monitoring system, consisting of an in vivo split GFP assay for activity-independent target detection and an mCherry-based secretion stress biosensor. The split GFP assay is based on the fusion of a target protein to the eleventh ß-sheet of sfGFP, which can complement a truncated sfGFP that lacks this ß-sheet named GFP1-10. The secretion stress biosensor makes use of the CssRS two component quality control system, which upregulates expression of mCherry in the htrA locus thereby allowing a fluorescence readout of secretion stress. RESULTS: The biosensor strain B. subtilis PAL5 was successfully constructed by exchanging the protease encoding gene htrA with mCherry via CRISPR/Cas9. The Fusarium solani pisi cutinase Cut fused to the GFP11 tag (Cut11) was used as a model enzyme to determine the stress response upon secretion mediated by signal peptides SPPel, SPEpr and SPBsn obtained from naturally secreted proteins of B. subtilis. An in vivo split GFP assay was developed, where purified GFP1-10 is added to the culture broth. By combining both methods, an activity-independent high-throughput method was created, that allowed optimization of Cut11 secretion. Using the split GFP-based detection assay, we demonstrated a good correlation between the amount of secreted cutinase and the enzymatic activity. Additionally, we screened a signal peptide library and identified new signal peptide variants that led to improved secretion while maintaining low stress levels. CONCLUSION: Our results demonstrate that the combination of a split GFP-based detection assay for secreted proteins with a secretion stress biosensor strain enables both, online detection of extracellular target proteins and identification of bottlenecks during protein secretion in B. subtilis. In general, the system described here will also enable to monitor the secretion stress response provoked by using inducible promoters governing the expression of different enzymes.

Sleep consolidates stimulus-response learning.

Performing a motor response to a sensory stimulus creates a memory trace whose behavioral correlates are classically investigated in terms of repetition priming effects. Such stimulus-response learning entails two types of associations that are partly independent: (1) an association between the stimulus and the motor response and (2) an association between the stimulus and the classification task in which it is encountered. Here, we tested whether sleep supports long-lasting stimulus-response learning on a task requiring participants (1) for establishing stimulus-classification associations to classify presented objects along two different dimensions ("size" and "mechanical") and (2) as motor response (action) to respond with either the left or right index finger. Moreover, we examined whether strengthening of stimulus-classification associations is preferentially linked to nonrapid eye movement (non-REM) sleep and strengthening of stimulus-action associations to REM sleep. We tested 48 healthy volunteers in a between-subjects design comparing postlearning retention periods of nighttime sleep versus daytime wakefulness. At postretention testing, we found that sleep supports consolidation of both stimulus-action and stimulus-classification associations, as indicated by increased reaction times in "switch conditions"; that is, when, at test, the acutely instructed classification task and/or correct motor response for a given stimulus differed from that during original learning. Polysomnographic recordings revealed that both kinds of associations were correlated with non-REM spindle activity. Our results do not support the view of differential roles for non-REM and REM sleep in the consolidation of stimulus-classification and stimulus-action associations, respectively.

Improved awareness of physical activities is associated with a gain of fitness and a stable body weight in breast cancer patients during the first year of antineoplastic therapy: the BEGYN-1 study.

Background: Breast cancer is the most frequent cancer in women. Reduced physical activity and overweight are associated with poor prognosis. Breast cancer patients have a high risk to gain weight, lose muscle mass and reduce physical activity during therapy. Concepts are urgently needed to motivate patients to engage in physical activity. Methods: 110 non-metastatic breast cancer patients were included in the prospective observational BEGYN-1 study. Physiological parameters and body composition were measured before the start of therapy and then quarterly for one year. Patients used a fitness tracker and documented their physical activity in a diary throughout the study. Results: Although the patients were not offered any guided exercise, and despite the restrictions during the COVID-19 pandemic, they increased their physical activity (metabolic equivalent of task (MET) -minutes): p<0.001), physical fitness (decreasing resting heart rate: p=0.001) and did not gain weight (median - 0.4kg) over the course of the study. Conclusion: Improved awareness of physical activity is associated with an increase in physical activity, fitness, and a stable weight during the first year of therapy in breast cancer patients. Counselling at diagnosis should motivate patients to engage in physical activity, wear a fitness tracker and document activities.

Data-driven tailoring of molecular dipole polarizability and frontier orbital energies in chemical compound space.

Understanding correlations - or lack thereof - between molecular properties is crucial for enabling fast and accurate molecular design strategies. In this contribution, we explore the relation between two key quantities describing the electronic structure and chemical properties of molecular systems: the energy gap between the frontier orbitals and the dipole polarizability. Based on the recently introduced QM7-X dataset, augmented with accurate molecular polarizability calculations as well as analysis of functional group compositions, we show that polarizability and HOMO-LUMO gap are uncorrelated when considering sufficiently extended subsets of the chemical compound space. The relation between these two properties is further analyzed on specific examples of molecules with similar composition as well as homooligomers. Remarkably, the freedom brought by the lack of correlation between molecular polarizability and HOMO-LUMO gap enables the design of novel materials, as we demonstrate on the example of organic photodetector candidates.

Analysis of a Deep Learning Model for 12-Lead ECG Classification Reveals Learned Features Similar to Diagnostic Criteria.

Despite their remarkable performance, deep neural networks remain unadopted in clinical practice, which is considered to be partially due to their lack of explainability. In this work, we apply explainable attribution methods to a pre-trained deep neural network for abnormality classification in 12-lead electrocardiography to open this "black box" and understand the relationship between model prediction and learned features. We classify data from two public databases (CPSC 2018, PTB-XL) and the attribution methods assign a "relevance score" to each sample of the classified signals. This allows analyzing what the network learned during training, for which we propose quantitative methods: average relevance scores over a) classes, b) leads, and c) average beats. The analyses of relevance scores for atrial fibrillation and left bundle branch block compared to healthy controls show that their mean values a) increase with higher classification probability and correspond to false classifications when around zero, and b) correspond to clinical recommendations regarding which lead to consider. Furthermore, c) visible P-waves and concordant T-waves result in clearly negative relevance scores in atrial fibrillation and left bundle branch block classification, respectively. Results are similar across both databases despite differences in study population and hardware. In summary, our analysis suggests that the DNN learned features similar to cardiology textbook knowledge.

CDK4/6 Inhibitors in Advanced HR+/HER2 - Breast Cancer: A Multicenter Real-World Data Analysis.

Purpose: CDK4/6 inhibitors (CDK4/6i) combined with endocrine therapy are considered standard-of-care for first-line therapy of patients with hormone receptor positive, HER2 negative, advanced breast cancer (HR+/HER2- ABC). Superiority of combination therapy over endocrine monotherapy has been demonstrated in a multitude of randomized controlled trials (RCTs) in phase III and IV. However, RCTs reflect clinical reality only to a limited extent, as narrow inclusion criteria lead to a selected patient collective. Here, we present real-world data (RWD) on CDK4/6i treatment in patients with HR+/HER2- ABC at four certified German university breast cancer centers. Methods: Patients diagnosed with HR+/HER2- ABC who were treated in clinical routine with CDK4/6i between November 2016 and December 2020 at four certified German university breast cancer centers (Saarland University Medical Center, University Medical Center Charité Berlin, University Medical Center Bonn, and University Medical Center Hospital Schleswig-Holstein, Campus Kiel) were identified and enrolled in this retrospective study. Clinicopathological characteristics and clinical outcomes were recorded with particular emphasis on CDK4/6i therapy course [progression-free survival (PFS) following treatment initiation, toxicity, dose reduction, therapy discontinuation, prior and subsequent therapy line]. Results: Data from n = 448 patients were evaluated. The mean patient age was 63 (±12) years. Of these patients, n = 165 (36.8%) were primarily metastasized, and n = 283 (63.2%) had secondary metastatic disease. N = 319 patients (71.3%) received palbociclib, n = 114 patients (25.4%) received ribociclib, and n = 15 patients (3.3%) received abemaciclib, respectively. Dose reduction was performed in n = 132 cases (29.5%). N = 57 patients (12.7%) discontinued the treatment with CDK4/6i due to side effects. N = 196 patients (43.8%) experienced disease progression under CDK4/6i treatment. The median PFS was 17 months. Presence of hepatic metastases and prior therapy lines were associated with shorter PFS, whereas estrogen positivity and dose reduction due to toxicity were positively associated with PFS. Presence of bone and lung metastases, progesterone positivity, Ki67 index, grading, BRCA1/2 and PIK3CA mutation status, adjuvant endocrine resistance, and age did not significantly impact on PFS. Conclusion: Our RWD analysis on CDK4/6i treatment in Germany supports data from RCTs regarding both treatment efficacy and safety of CDK4/6i for treatment of patients with HR+/HER2- ABC. In comparison to data from the pivotal RCTs, median PFS was lower but within the expected range for RWD, which could result from inclusion of patients with more advanced diseases (i.e., higher therapy lines) to our dataset.

Prevalence and Relevance of Vitamin D Deficiency in Newly Diagnosed Breast Cancer Patients: A Pilot Study.

(1) Background: Vitamin D plays an important role in many types of cancer. It was the aim of this study to analyze serum 25-hydroxyvitamin D (25(OH)D) levels in newly diagnosed breast cancer patients, and the association with prognostic and lifestyle factors. (2) Methods: 110 non-metastatic breast cancer patients were included in the prospective observational "BEGYN" study at Saarland University Medical Center between September 2019 and January 2021. At the initiation visit, serum 25(OH)D levels were measured. Clinicopathological data on prognosis, nutrition, and lifestyle were extracted from data files and obtained using a questionnaire. (3) Results: Median serum 25(OH)D in breast cancer patients was 24 ng/mL (range 5-65 ng/mL), with 64.8% of patients being vitamin D deficient. 25(OH)D was higher among patients that reported the use of vitamin D supplements (43 ng/mL versus 22 ng/mL; p < 0.001), and in summer compared to other seasons (p = 0.03). Patients with moderate vitamin D deficiency were less likely to have triple negative breast cancer (p = 0.047). (4) Conclusions: Routinely measured vitamin D deficiency is common in breast cancer patients and needs to be detected and treated. However, our results do not support the hypothesis that vitamin D deficiency may be a main prognostic factor for breast cancer.

Explore or exploit? A model-based screening strategy for PETase secretion by Corynebacterium glutamicum.

Extracellular production of target proteins simplifies downstream processing due to obsolete cell disruption. However, optimal combinations of a heterologous protein, suitable signal peptide, and secretion host can currently not be predicted, resulting in large strain libraries that need to be tested. On the experimental side, this challenge can be tackled by miniaturization, parallelization, and automation, which provide high-throughput screening data. These data need to be condensed into a candidate ranking for decision-making to focus bioprocess development on the most promising candidates. We screened for Bacillus subtilis signal peptides mediating Sec secretion of two polyethylene terephthalate degrading enzymes (PETases), leaf-branch compost cutinase (LCC) and polyester hydrolase mutants, by Corynebacterium glutamicum. We developed a fully automated screening process and constructed an accompanying Bayesian statistical modeling framework, which we applied in screenings for highest activity in 4-nitrophenyl palmitate degradation. In contrast to classical evaluation methods, batch effects and biological errors are taken into account and their uncertainty is quantified. Within only two rounds of screening, the most suitable signal peptide was identified for each PETase. Results from LCC secretion in microliter-scale cultivation were shown to be scalable to laboratory-scale bioreactors. This work demonstrates an experiment-modeling loop that can accelerate early-stage screening in a way that experimental capacities are focused to the most promising strain candidates. Combined with high-throughput cloning, this paves the way for using large strain libraries of several hundreds of strains in a Design-Build-Test-Learn approach.