Machine learning-based surrogate model assisting stochastic model predictive control of urban drainage systems

Quantifying the uncertainty of stormwater inflow is critical for improving the resilience of urban drainage systems (UDSs). However, the high computational complexity and time consumption obstruct the implementation of uncertainty-addressing methods for real-time control of UDSs. To address this issue, this study developed a machine learning-based surrogate model (MLSM) that maintains high-fidelity descriptions of drainage dynamics and meanwhile diminishes the computational complexity. With stormwater inflow and controls as inputs and system overflow as the output, MLSM is able to fast evaluate system performance, and therefore stochastic optimization becomes feasible. Thus, a real-time control strategy was built by combining MLSM with the stochastic model predictive control. This strategy used stochastic stormwater inflow scenarios as input and aimed to minimize the expected overflow under all scenarios. An ensemble of stormwater inflow scenarios was generated by assuming the forecast errors follow normal distributions. To downsize the ensemble, representative scenarios with their probabilities were selected using the simultaneous backward reduction method. The proposed control strategy was applied to a combined UDS of China. Results are as follows. (1) MLSM fit well with the original high-fidelity urban drainage model, while the computational time was reduced by 99.1%. (2) The proposed strategy consistently outperformed the classical deterministic model predictive control in both magnitude and duration dimensions of system resilience, when the consumed time compatible is with the real-time operation. It is indicated that the proposed control strategy could be used to inform the real-time operation of complex UDSs and thus enhance system resilience to uncertainty.

Variations of olfactory function with circadian timing and chronotype

BACKGROUND

Cumulative animal studies have suggested that olfaction can be regulated by circadian clock. However, human studies on the topic are relatively limited. The present study thus aimed to investigate diurnal variation in olfaction in healthy adults while examining potential modulating factors.

METHODS

We conducted four rounds of testing on 56 healthy adults (32 women) aged 31 ± 12 years, throughout a single day, during morning (8:00-10:00 h), noon (12:00-14:00 h), afternoon (16:00-18:00 h), and evening (20:00-22:00 h). At the first appointment, participants completed full olfactory function testing using the Sniffin’ Sticks, questionnaires on medical history, nasal symptoms, sleep quality, and chronotype, and were assessed for blood pressure, heart rate, peak nasal inspiratory flow (PNIF), attention level, and rated their smell ability, nasal patency, wakefulness, and concentration level using visual analog scale (VAS) ratings. Subsequent appointments measured olfactory threshold, attentional level, PNIF, blood pressure, heart rate and VAS ratings repeatedly.

RESULTS

Olfactory threshold (OT) scores varied significantly between different times of the day, with the highest score in the evening and the lowest in the morning. Similar differences were also observed in PNIF, with the highest value in the evening and the lowest in the morning. However, there were no significant correlations between OT score and PNIF across all four-time testing, as well as between differences in [OT evening â€" OT morning] and [PNIF evening â€" PNIF morning]. Furthermore, a generalized linear mixed model indicated that the testing time of the morning, evening chronotype, self-reported body mass index (BMI), rated smell ability, and rated nasal patency significantly predicted the Sniffin' Sticks OT score.

CONCLUSIONS

Olfactory function fluctuates throughout the waking hours of the day, with the highest olfactory sensitivity observed in the evening and the lowest in the morning. This pattern is also seen in nasal patency. However, it appears that the circadian changes of nasal airflow may not significantly depend on the circadian changes of the olfactory sensitivity. In addition, chronotype and BMI may regulate such olfactory-circadian variation. These findings provide important insights for future research on the accurate diagnosis and treatment of olfactory dysfunction.

Native autoantigen complex detects pemphigoid autoantibodies

Pemphigoid diseases are a group of autoimmune disorders characterized by subepidermal blistering in the skin and mucosa. Among them, mucous membrane pemphigoid (MMP) autoantibodies are characterized by targeting multiple molecules in the hemidesmosomes, including collagen XVII, laminin-332, and integrin a6/β4. Traditionally, recombinant proteins of the autoantigens have been employed to identify circulating autoantibodies by immune assays. However, developing an efficient detection system for MMP autoantibodies has been challenging because the autoantibodies have heterogeneous profiles and the antibody titers are typically low. In this study, we introduce an ELISA that takes advantage of a native autoantigen complex rather than simple recombinant proteins. We generated HaCaT keratinocytes with a DDDDK-tag knocked in at the COL17A1 locus by CRISPR/Cas9-mediated gene editing. Immunoprecipitation using the DDDDK-tag isolated a native complex that contained full-length and processed collagen XVII and integrin α6/β4. Then, we used the complex proteins to prepare an ELISA system and enrolled 55 MMP cases to validate its diagnostic performance. The sensitivity and specificity of the ELISA for detecting MMP autoantibodies were 70.9% and 86.7%, respectively, far superior to those of conventional assays. In autoimmune diseases such as MMP, in which autoantibodies target various molecules, isolating the antigen-protein complexes can help establish a diagnostic system.

AB0227 TREATMENT SEQUENCING PATTERNS AND COMPARATIVE EFFICACY IN PATIENTS WITH RHEUMATOID ARTHRITIS FROM A REAL-WORLD SETTING

Background

The European League Against Rheumatism (EULAR)1 recently provided updated guidelines regarding the initiation and modification of disease-modifying antirheumatic drug (DMARD) therapy in patients with Rheumatoid Arthritis (RA). Therefore, real-world evidence studies are warranted to provide insights into first-line DMARD utilization and durability of response in the second-line setting.

Objectives

To analyze RA treatment patterns in real-world data and compare durability of response between second-line DMARDs + anti-TNF (TNFi) therapies vs. TNFi monotherapy.

Methods

Electronic health records (EHRs) from a large health system in the Northeast US were used to identify RA patients. Lines of therapy were defined based on confirmed prescriptions for DMARDs and TNFi therapies. Time to next treatment (TTNT) was the primary outcome to estimate durability of response. Time-to-event analyses were performed using Kaplan-Meier and log-rank test methods. In addition, a Cox Proportional-Hazards (CoxPH) model was used to evaluate covariates as independent predictors of disease progression.

Results

Our study cohort consisted of 8,040 patients who had at least one line of therapy for RA. Conventional synthetic DMARDs (csDMARDs) were the predominant first line of therapy in this dataset (71.3%), followed by TNFi alone (11.1%) or TNFi combined with csDMARD (9.1%) (Figure 1).

For patients who had csDMARD as their first line of therapy, 22.93% progressed to second line treatment. Among them 36.2% patients were TNFi with or without in combination with csDMARDs. In the second-line, TNFi + csDMARDs were associated with a longer TTNT (median time: 13.1 months vs 6.1 months, P < 0.005) compared to TNFi monotherapy. The multiple variable CoxPH model (adjusted for age, gender, and race) demonstrated that second-line TNFi + csDMARDs had a lower hazard rate when compared to TNFi monotherapy (HR = 0.74, 95% CI: 0.36 - 1.12, p < 0.005).

Conclusion

We demonstrated the first comprehensive treatment sequencing patterns in RA from a real-world setting. As a second-line therapy for patients with inadequate response to csDMARDS, the TNFi + csDMARDs combination may improve duration of response when compared to TNFi monotherapy. Results from this study will inform future sequencing strategies to improve patient outcomes.

References

[1]Smolen, Josef S., Robert B. M. Landewé, Johannes W. J. Bijlsma, Gerd R. Burmester, Maxime Dougados, Andreas Kerschbaumer, Iain B. McInnes, et al. 2020. “EULAR Recommendations for the Management of Rheumatoid Arthritis with Synthetic and Biological Disease-Modifying Antirheumatic Drugs: 2019 Update.” Annals of the Rheumatic Diseases 79 (6): 685–99.

Disclosure of Interests

Lei Ai: None declared, Mitchell Higashi: None declared, Kyeryoung Lee: None declared, Zongzhi Liu: None declared, Lan Jin: None declared, Kalpana Raja: None declared, Yun Mai: None declared, Tomi Jun: None declared, William Oh Consultant of: Janssen

Pfizer, Aviva Beckmann: None declared, Emilio Schadt: None declared, Zachary Schadt: None declared, Rick Wallsten: None declared, Ediz Calay: None declared, Andrew Kasarskis: None declared, Qi Pan: None declared, Eric Schadt Speakers bureau: Eli Lilly, Consultant of: SAB of Eli Lilly

Celgene, Xiaoyan Wang: None declared

Identification of hematopoietic cell populations from the dermal papillae of human hair follicles

Hematopoietic stem cell (HSC) activity has been identified from the hair follicles (HFs) in mice; however, it has not been identified in human HFs. We used immunohistochemistry and flow cytometry to identify cultured dermal papilla (DP) cells expressing CD45 to test for hematopoietic activity in colony-forming assays of granulocyte/macrophage hematopoietic progenitors (CFU-GM). Occasional CD45-positive cells were detected in cultured DP cells. After in vitro stimulation with IL-3, granulocyte-macrophage colony-stimulating factor (GM-CSF) and granulocyte colony-stimulating factor (G-CSF) for 7 days, about 1% of the cells were CD45-positive by flow cytometry analysis, an fifty-fold expansion in cell numbers. We further examined whether mesenchymal stem/progenitor cells reside in human dermal papillae. Cultured DP papilla cells incubated with monoclonal antibodies to remove the CD45 positive cells were induced into multilineage differentiation with the formation of CFU-GM. Our findings preliminarily indicate that human dermal papilla contain at least a CD45-positive hematopoietic cell population and a mesenchymal stem/progenitor cell population.

Thymocytes and cultured thymic epithelial cells express transcripts encoding alpha-3, alpha-5 and beta-4 subunits of neuronal nicotinic acetylcholine receptors: preferential transcription of the alpha-3 and beta-4 genes by immature CD4 + 8 + thymocytes

Thymic tissues express transcripts encoding the alpha-3, alpha-5 and beta-4 subunits of nicotinic neuronal acetylcholine receptors (AcChRs) suggesting that neuronal AcChRs similar to those expressed in ganglia are expressed in the thymus. Transcription occurs in both isolated thymocytes and thymic epithelial cells. RT-PCR analyses of thymocyte subsets indicate that immature CD4 + 8 + thymocytes express higher levels of the alpha-3 and beta-4 transcripts than more mature thymocytes. Compared to freshly isolated thymocytes, peripheral blood lymphocytes do not express alpha-3 and beta-4 AcChR subunit transcripts. Cultured thymocytes rapidly down-regulate transcription of the alpha-3 and beta-4 AcChR subunit genes by a process that is not reversed by stimulation with phytohemagglutinin and IL-2. Thus our results indicate that there is transcriptional regulation of neuronal AcChR subunit genes during the process of thymocyte maturation and that factors within the thymic microenvironment influence expression of the alpha-3 and beta-4 AcChR subunit genes by developing T cells.

Splicing of an anti-sense Alu sequence generates a coding sequence variant for the alpha-3 subunit of a neuronal acetylcholine receptor

In this report we demonstrate that an alpha-3 acetylcholine receptor subunit transcriptional variant originates through alternative splicing of a complementary sequence of the right arm of an Alu element. This element is located within the 5.1 Kb intron found between exons 5 and 6 of the alpha-3 acetylcholine receptor subunit gene. The transcriptional variant originates from the normal splicing process and carries an in-frame stop codon. If translated, it should encode for a peptide lacking the 4th transmembrane domain of the normal subunit.