OsNAC3 regulates seed germination involving abscisic acid pathway and cell elongation in rice.

Seed germination is a critical trait for the success of direct seeding in rice cultivation. However, the underlying mechanism determining seed germination is largely unknown in rice. Here, we report that NAC transcription factor OsNAC3 positively regulates seed germination of rice. OsNAC3 regulates seed germination involving abscisic acid (ABA) pathway and cell elongation. OsNAC3 can directly bind to the promoter of ABA catabolic gene OsABA8ox1 and cell expansion gene OsEXP4, which consequently activates their expressions during seed germination. We also find that the expression of OsEXP4 is reduced by ABA during seed germination in rice. OsNAC3 regulates seed germination by influencing cell elongation of the embryo through directly affecting OsEXP4 expression and indirectly ABA-medicated OsEXP4 expression. The OsNAC3 elite haplotype is useful for genetic improvement of seed germination, and overexpression of OsNAC3 can significantly increase seed germination. We therefore propose that OsNAC3 is a potential target in breeding of rice varieties with high seed germination for direct seeding cultivation.

Machine learning-based pathomics signature of histology slides as a novel prognostic indicator in primary central nervous system lymphoma.

PURPOSE: To develop and validate a pathomics signature for predicting the outcomes of Primary Central Nervous System Lymphoma (PCNSL). METHODS: In this study, 132 whole-slide images (WSIs) of 114 patients with PCNSL were enrolled. Quantitative features of hematoxylin and eosin (H&E) stained slides were extracted using CellProfiler. A pathomics signature was established and validated. Cox regression analysis, receiver operating characteristic (ROC) curves, Calibration, decision curve analysis (DCA), and net reclassification improvement (NRI) were performed to assess the significance and performance. RESULTS: In total, 802 features were extracted using a fully automated pipeline. Six machine-learning classifiers demonstrated high accuracy in distinguishing malignant neoplasms. The pathomics signature remained a significant factor of overall survival (OS) and progression-free survival (PFS) in the training cohort (OS: HR 7.423, p < 0.001; PFS: HR 2.143, p = 0.022) and independent validation cohort (OS: HR 4.204, p = 0.017; PFS: HR 3.243, p = 0.005). A significantly lower response rate to initial treatment was found in high Path-score group (19/35, 54.29%) as compared to patients in the low Path-score group (16/70, 22.86%; p < 0.001). The DCA and NRI analyses confirmed that the nomogram showed incremental performance compared with existing models. The ROC curve demonstrated a relatively sensitive and specific profile for the nomogram (1-, 2-, and 3-year AUC = 0.862, 0.932, and 0.927, respectively). CONCLUSION: As a novel, non-invasive, and convenient approach, the newly developed pathomics signature is a powerful predictor of OS and PFS in PCNSL and might be a potential predictive indicator for therapeutic response.

MPLC-based low-modal-crosstalk nine-mode-group demultiplexer for DSP-free IM/DD MDM transmission over MMF.

Weakly coupled mode-division multiplexing (MDM) transmission over legacy laid multimode fiber (MMF) has great economic efficiency and can enormously enhance the capacity of short-reach optical interconnections. In order to be compatible with cost-efficient intensity-modulation/direct-detection (IM/DD) transceivers, weakly coupled mode-group demultiplexers that can simultaneously receive each mode group of MMFs are highly desired. In this paper, we propose a scalable low-modal-crosstalk mode-group demultiplexer over MMF based on multiplane light conversion (MPLC). Multiple input Hermite-Gaussian (HG) modes of MMF are first converted to bridging modes that are composed of H G 00 modes distributed as a right-angle triangle in Cartesian coordinates, and then each H G 00 mode belonging to a degenerate mode group is mapped to different overlapped H G n0 modes with vertical orientation for simultaneous detection. With the help of bridging modes, the MPLC-based mode-group demultiplexer can efficiently demultiplex all mode groups in standard MMFs with less than 20 phase masks. A nine-mode-group demultiplexer is further designed for demonstration, and simulation results show that the MPLC-based demultiplexer achieves low modal crosstalk of lower than -22.3d B at 1550 nm and lower than -17.9d B over the C-band for all the nine mode groups with only 16 phase masks.

Low-CD weakly-coupled FMF for short-reach IM/DD MDM transmission.

Weakly-coupled mode division multiplexing (MDM) technique is a promising candidate for capacity enhancement of short-reach optical interconnections, for which the multiple-ring-core few-mode fiber (MRC-FMF) has been proven to be an effective design method to suppress distributed modal crosstalk. Similar to low chromatic-dispersion (CD) O-band transmission based on single-mode fibers (SMF), all the mode channels in a weakly-coupled FMF for short-reach applications should achieve low CD to support intensity-modulation/direct-detection (IM/DD) transmission. In this paper, we propose, for the first time to the best of our knowledge, an index perturbation method to adjust both effective index and CD of each mode in an MRC-FMF. Firstly, an accurate modeling method to model the relationship between SiO2-GeO2 material index and the germanium concentration at different wavelengths is proposed by analyzing the index characteristics of 4 kinds of germanium-doped fused silica SMFs at the same fabrication processing, which could be utilized to calculate the CD characteristics for an MRC-FMF with perturbed index profile. Then, based on the perturbation method considering the influences on both effective index and CD, a weakly-coupled low-CD MRC-FMF supporting 4 linearly-polarized (LP) modes is designed and fabricated. The measured minimum effective index difference min|Δneff| among all modes is larger than 1.3 × 10-3, and the CD values of all the modes lie between -6 and +6 ps/km/nm ranging from 1280 to 1320 nm, which agree well with the design. The 2-km transmission experiment indicates that the fabricated MRC-FMF could support stable digital-signal-processing (DSP)-free IM/DD transmission for all the 4 LP modes. This work is beneficial to the application of short-reach weakly-coupled MDM systems.

Low-modal-crosstalk orthogonal combine reception for degenerate modes in IM/DD MDM transmission.

Weakly-coupled mode division multiplexing (MDM) techniques supporting intensity modulation and direct detection (IM/DD) transmission is a promising candidate to enhance the capacity of short-reach applications such as optical interconnections, in which low-modal-crosstalk mode multiplexers/demultiplexers (MMUX/MDEMUX) are highly desired. In this paper, we firstly propose an all-fiber low-modal-crosstalk orthogonal combine reception scheme for degenerate linearly-polarized (LP) modes, in which signals in both degenerate modes are firstly demultiplexed into the LP01 mode of single-mode fibers, and then are multiplexed into mutually orthogonal LP01 and LP11 modes of a two-mode fiber for simultaneous detection. Then a pair of 4-LP-mode MMUX/MDEMUX consisting of cascaded mode-selective couplers and orthogonal combiners are fabricated with side-polishing processing, which achieve low back-to-back modal crosstalk of lower than -18.51 dB and insertion loss of lower than 3.81 dB for all the 4 modes. Finally, a stable real-time 4 modes × 4λ × 10 Gb/s MDM-wavelength division multiplexing (WDM) transmission over 20-km few-mode fiber is experimentally demonstrated. The proposed scheme is scalable to support more modes and can pave the way to practical implementation of IM/DD MDM transmission applications.

A genome-wide association study reveals that the 2-oxoglutarate/malate translocator mediates seed vigor in rice.

Seed vigor is an important trait for the direct seeding of rice (Oryza sativa L.). In this study, we examined the genetic architecture of variation in the germination rate using a diverse panel of rice accessions. Four quantitative trait loci for germination rate were identified using a genome-wide association study during early germination. One candidate gene, encoding the 2-oxoglutarate/malate translocator (OsOMT), was validated for qGR11. Disruption of this gene (Osomt mutants) reduced seed vigor, including seed germination and seedling growth, in rice. Functional analysis revealed that OsOMT influences seed vigor mainly by modulating amino acid levels and glycolysis and tricarboxylic acid cycle processes. The levels of most amino acids, including the Glu family (Glu, Pro, Arg, and GABA), Asp family (Asp, Thr, Lys, Ile, and Met), Ser family (Ser, Gly, and Cys), and others (His, Ala, Leu, and Val), were significantly reduced in the mature grains and the early germinating seeds of Osomt mutants compared to wild type (WT). The glucose and soluble sugar contents, as well as adenosine triphosphate levels, were significantly decreased in germinating seeds of Osomt mutants compared to WT. These results provide important insights into the role of OsOMT in seed vigor in rice.

Artificial Intelligence-Aided Multiple Tumor Detection Method Based on Immunohistochemistry-Enhanced Dark-Field Imaging.

The immunohistochemical method serves as one of the most practical tools in clinical cancer detection and thus has great application value to overcome the existing limits of the conventional method and further improve the detecting efficiency and sensitivity. This study employed 3,3'-diaminobenzidine (DAB), a conventional color indicator for immunohistochemistry, as a novel high-sensitive scattering reagent to provide a multidimensional image signal varying with the overexpression rate of tumor markers. Based on the scattering properties of DAB aggregates, an efficient and robust artificial intelligence-aided immunohistochemical method based on dark-field imaging has been established, with improvement in both the imaging quality and interpretation efficiency in comparison with the conventional manual-operated immunohistochemical method. Referencing the diagnosis from three independent pathologists, this method succeeded in detecting HER2 overexpressed breast tumors with a sensitivity of 95.2% and a specificity of 100.0%; meanwhile, it was found to be applicable for non-small-cell lung tumors and malignant lymphoma as well. As demonstrated, this study provided an effective and reliable means for making diagnostic suggestions, which exhibited great potential in multiple tumor pathological detection at low cost.

Genome-wide association study reveals that the cupin domain protein OsCDP3.10 regulates seed vigour in rice.

Seed vigour is an imperative trait for the direct seeding of rice. In this study, we examined the genetic regulation of seedling percentage at the early germination using a genome-wide association study in rice. One major quantitative trait loci qSP3 for seedling percentage was identified, and the candidate gene was validated as qSP3, encoding a cupin domain protein OsCDP3.10 for the synthesis of 52 kDa globulin. Disruption of this gene in Oscdp3.10 mutants reduced the seed vigour, including the germination potential and seedling percentage, at the early germination in rice. The lacking accumulation of 52 kDa globulin was observed in the mature grains of the Oscdp3.10 mutants. The significantly lower amino acid contents were observed in the mature grains and the early germinating seeds of the Oscdp3.10 mutants compared with those of wild-type. Rice OsCDP3.10 regulated seed vigour mainly via modulating the amino acids e.g. Met, Glu, His, and Tyr that contribute to hydrogen peroxide (H2 O2 ) accumulation in the germinating seeds. These results provide important insights into the application of seed priming with the amino acids and the selection of OsCDP3.10 to improve seed vigour in rice.

A single nucleotide polymorphism-based formula to predict the risk of propofol TCI concentration being over 4 µg mL-1 at the time of loss of consciousness.

We aim to develop a formula based on single nucleotide polymorphisms (SNPs) to predict whether the propofol target-controlled infusion (TCI) concentration would be over 4 µg mL-1 at the time of loss of consciousness (LOC). We recruited 184 patients undergoing thyroid or breast surgeries with propofol anaesthesia. A total of 48 SNPs of CYP2B6, CYP2C9, UGT1A9, HNF4A, ABCB1, ABCC4, ABCG2, GABRA2, GABRA4, GABRB1, GABRB3, GABRG2, GABBR2, GAD1, SLC1A3, BDNF, and NRXN1, previously associated with propofol metabolic and pharmacology pathway, were genotyped. The formula was developed in the training cohort using the least absolute shrinkage and selection operator logistic regression model, and then validated in the testing cohort. The SNPs, GABBR2 rs1167768, GABBR2 rs1571927, NRXN1 rs601010, BDNF rs2049046, GABRA4 rs1512135, UGT1A9 rs11692021, GABBR2 rs2808536, HNF4A rs1884613, GABRB3 rs2017247, and CYP2B6 rs3181842 were selected to construct the SNP-based formula, which was used to calculate the risk score for over 4 µg mL-1 TCI concentration of propofol at the time of LOC. Patients in the high-risk group were more likely to require a propofol concentration higher than 4 µg mL-1 and presented a longer LOC latency. The SNP-based formula may significantly improve the safety and effectiveness of propofol-induced anaesthesia.

Low-crosstalk laser-direct-writing FI/FO device for 8×100-Gbps optical interconnection.

Fan-in/fan-out (FI/FO) device with low crosstalk is essential for weakly coupled short-reach optical interconnect based on multicore fibers (MCF), for which the laser-direct-writing (LDW) technique is one of the preferred fabrication schemes. In this paper, the influence of FI/FO crosstalk on short-reach intensity-modulation/direction-detection MCF optical interconnection is firstly evaluated, and the crosstalk related to different refractive-index profiles of waveguides and misalignment is analyzed for LDW-FI/FO devices. Then low-crosstalk compact LDW-FI/FO devices matching 8-core MCF are fabricated, adopting multiple-scan method for waveguides with a flat-top refractive-index profile and aberration correction method for precise alignment. Owing to the low crosstalk, 8×100-Gbps optical interconnection over 10-km MCF is experimentally demonstrated with only 0.5-dB penalty compared to 10-km G.652D single-mode fiber transmission. Simulation results indicate that the transmission reach can be further extended to over 40 km. The proposed prototype system with low crosstalk is promising for high-speed optical interconnection applications.

Long-haul intermodal-MIMO-free MDM transmission based on a weakly coupled multiple-ring-core few-mode fiber.

Mode-division multiplexing (MDM) technique based on few-mode fibers (FMFs) can achieve multiplicative growth in single-fiber capacity by using different linearly polarized (LP) modes or mode groups as spatial channels. However, its deployment is seriously impeded because multiple-input multiple-output digital signal processing (MIMO-DSP) with huge computational load must be adopted to combat intermodal crosstalk for long-haul FMF transmission. In this paper, we present an intermodal-MIMO-free MDM transmission scheme based on weakly coupled multiple-ring-core FMF, which achieves ultralow distributed modal crosstalk (DMC) so that the signal in each LP mode can be independently received by single-LP-mode MIMO-DSP even after hundreds-of-kilometer transmission. Evaluation method for the required DMC levels is proposed and different transmission reaches are investigated by simulation. By adopting an improved method for quantitative DMC measurement, we show that the required DMC level for long-haul transmission is feasible. Finally, we experimentally demonstrate 1800-km LP01/LP02 multiplexed transmission and 525-km LP01/LP21/LP02 multiplexed transmission only adopting 2×2 or 4×4 MIMO-DSP. The proposed scheme may pave the way to practical applications of long-haul MDM techniques for the first time.

Distributed vibration sensor based on mode coupling in weakly coupled few-mode fibers.

In recent years, optical fiber distributed vibration sensors (DVSs) have received extensive investigation and play a significant role in different applications, such as structural health monitoring. In this Letter, we propose for the first time, to the best of our knowledge, a DVS mechanism based on linearly polarized mode coupling in weakly coupled few-mode fibers (FMFs), in which dynamic transverse stress induced by external vibration is measured with quantifiable and spatially resolvable mode coupling along the sensing FMF with ultralow inherent modal crosstalk. A swept-wavelength interferometer method is implemented and the involved data processing method is designed. A proof-of-concept DVS system is established and 5 Hz to 49 kHz frequency response, -50 dB detection sensitivity, and 22 m spatial resolution are successfully demonstrated based on a 9.6 km weakly coupled two-mode fiber. The wide frequency response over a long sensing length for the proposed scheme may extend the application range of DVS systems.

A genome-wide association study reveals that the glucosyltransferase OsIAGLU regulates root growth in rice.

Good root growth in the early post-germination stages is an important trait for direct seeding in rice, but its genetic control is poorly understood. In this study, we examined the genetic architecture of variation in primary root length using a diverse panel of 178 accessions. Four QTLs for root length (qRL3, qRL6, qRL7, and qRL11) were identified using genome-wide association studies. One candidate gene was validated for the major QTL qRL11, namely the glucosyltransferase OsIAGLU. Disruption of this gene in Osiaglu mutants reduced the primary root length and the numbers of lateral and crown roots. The natural allelic variations of OsIAGLU contributing to root growth were identified. Functional analysis revealed that OsIAGLU regulates root growth mainly via modulating multiple hormones in the roots, including levels of auxin, jasmonic acid, abscisic acid, and cytokinin. OsIAGLU also influences the expression of multiple hormone-related genes associated with root growth. The regulation of root growth through multiple hormone pathways by OsIAGLU makes it a potential target for future rice breeding for crop improvement.

A genome-wide association study reveals that the cytochrome b5 involved in seed reserve mobilization during seed germination in rice.

KEY MESSAGE: A candidate gene cytochrome b5 for the major QTL qSRMP9 for rice seed reserve mobilization was validated during seed germination using a genome-wide association study approach. Seed reserve mobilization plays important roles in the early seedling growth in rice. However, the genetic basis underlying this process is poorly understood. In this study, the genetic architecture of variation in seed reserve mobilization during seed germination was studied using a genome-wide association study approach in rice. Three quantitative trait loci (QTL) including qSRMP6, qSRMP9, and qSRMP12 for seed reserve mobilization percentage were identified. In which, the candidate gene cytochrome b5 (OsCyb5) for the major QTL qSRMP9 was validated. Disruption of this gene in Oscyb5 mutants reduced the seed reserve mobilization and seedling growth compared with wild-type (WT) in rice. There were no significant differences of grain size, starch, protein and total soluble sugar content in the mature grains between Oscyb5 mutants and WT. However, the α-amylase activity in the germinating seeds of Oscyb5 mutants was significantly decreased compared to that of WT, and then, the starch and sugar mobilization and the glucose accumulation during seed germination were significantly decreased in Oscyb5 mutants. Two elite haplotypes of OsCyb5 associated with the higher seed reserve mobilization percentage and its elite single nucleotide polymorphism variations were mainly existed in the INDICA and AUS accessions. The natural variation of OsCyb5 contributing to seed reserve mobilization might be useful for the future rice breeding.

The Rice Small Auxin-Up RNA Gene OsSAUR33 Regulates Seed Vigor via Sugar Pathway during Early Seed Germination.

Seed vigor affects seed germination and seedling emergence, and therefore is an important agronomic trait in rice. Small auxin-up RNAs (SAURs) function in a range of developmental processes, but their role in seed vigor remains unclear. Here, we observed that disruption of OsSAUR33 resulted in reduced germination rates and low seed uniformity in early germination. Expression of OsSAUR33 was higher in mature grains and early germinating seeds. RNA-seq analysis revealed that OsSAUR33 modulated seed vigor by affecting the mobilization of stored reserves during germination. Disruption of OsSAUR33 increased the soluble sugar content in dry mature grains and seeds during early germination. OsSAUR33 interacted with the sucrose non-fermenting-1-related protein kinase OsSnRK1A, a regulator of the sugar signaling pathway, which influences the expression of sugar signaling-related genes during germination. Disruption of OsSAUR33 increased sugar-sensitive phenotypes in early germination, suggesting OsSAUR33 likely affects seed vigor through the sugar pathway. One elite haplotype of OsSAUR33 associated with higher seed vigor was identified mainly in indica accessions. This study provides insight into the effects of OsSAUR33 on seed vigor in rice.

A quantitative trait locus, qSE3, promotes seed germination and seedling establishment under salinity stress in rice.

Seed germination is a complex trait determined by both quantitative trait loci (QTLs) and environmental factors and also their interactions. In this study, we mapped one major QTLqSE3 for seed germination and seedling establishment under salinity stress in rice. To understand the molecular basis of this QTL, we isolated qSE3 by map-based cloning and found that it encodes a K+ transporter gene, OsHAK21. The expression of qSE3 was significantly upregulated by salinity stress in germinating seeds. Physiological analysis suggested that qSE3 significantly increased K+ and Na+ uptake in germinating seeds under salinity stress, resulting in increased abscisic acid (ABA) biosynthesis and activated ABA signaling responses. Furthermore, qSE3 significantly decreased the H2 O2 level in germinating seeds under salinity stress. All of these seed physiological changes modulated by qSE3 might contribute to seed germination and seedling establishment under salinity stress. Based on analysis of single-nucleotide polymorphism data of rice accessions, we identified a HAP3 haplotype of qSE3 that was positively correlated with seed germination under salinity stress. This study provides important insights into the roles of qSE3 in seed germination and seedling establishment under salinity stress and facilitates the practical use of qSE3 in rice breeding.

Indole-3-acetate beta-glucosyltransferase OsIAGLU regulates seed vigour through mediating crosstalk between auxin and abscisic acid in rice.

Seed vigour is an important trait for direct seeding in rice. In this study, indole-3-acetate beta-glucosyltransferase OsIAGLU was cloned in rice, and its roles on seed vigour were mainly investigated. Disruption of OsIAGLU resulted in low seed vigour in rice. Quantitative RT-PCR analysis showed that the expressions of OsIAGLU were relatively higher in the late developing and the early germinating seeds and were significantly induced by indole-3-acetic acid (IAA) and abscisic acid (ABA). Transcriptome analysis revealed that the IAA- and ABA-related genes were involved in the OsIAGLU regulation of seed vigour in rice. The higher levels of free IAA and ABA were identified in germinating seeds of osiaglu mutants compared to wild-type (WT) plants. When treated with exogenous IAA and ABA, the osiaglu mutants and WT plants showed sensitivity to ABA while not IAA, but the exogenous IAA amplified ABA-induced reduction of seed vigour in rice. The continuously higher expressions of ABA-INSENSITIVE 3 (OsABI3) and OsABI5 occurred in germinating seeds of osiaglu mutants compared to WT plants. The regulation of seed vigour by OsIAGLU might be through modulating IAA and ABA levels to alert OsABIs expression in germinating seeds in rice. Based on analysis of single-nucleotide polymorphism data of rice accessions, two haplotypes of OsIAGLU that positively correlated with seed vigour were identified in indica accessions. This study provides important insights into the roles of OsIAGLU on seed vigour and facilitates the practical use of OsIAGLU in rice breeding.

Influence of isopropylmalate synthase OsIPMS1 on seed vigour associated with amino acid and energy metabolism in rice.

Seed vigour is an imperative trait for the direct seeding of rice. Isopropylmalate synthase (IPMS) catalyses the committed step of leucine (Leu) biosynthesis, but its effect on seed vigour remains unclear. In this study, rice OsIPMS1 and OsIPMS2 was cloned, and the roles of OsIPMS1 in seed vigour were mainly investigated. OsIPMS1 and OsIPMS2 catalyse Leu biosynthesis, and Leu feedback inhibits their IPMS activities. Disruption of OsIPMS1 resulted in low seed vigour under various conditions, which might be tightly associated with the reduction of amino acids in germinating seeds. Eleven amino acids that associated with stress tolerance, GA biosynthesis and tricarboxylic acid (TCA) cycle were significantly reduced in osipms1 mutants compared with those in wide type (WT) during seed germination. Transcriptome analysis indicated that a total of 1209 differentially expressed genes (DEGs) were altered in osipms1a mutant compared with WT at the early germination stage, wherein most of the genes were involved in glycolysis/gluconeogenesis, protein processing, pyruvate, carbon, fructose and mannose metabolism. Further analysis confirmed that the regulation of OsIPMS1 in seed vigour involved in starch hydrolysis, glycolytic activity and energy levels in germinating seeds. The effects of seed priming were tightly associated with the mRNA levels of OsIPMS1 in priming seeds. The OsIPMS1 might be used as a biomarker to determine the best stop time-point of seed priming in rice. This study provides novel insights into the function of OsIPMS1 on seed vigour and should have practical applications in seed priming of rice.

Prototype system for real-time IM/DD MDM transmission based on multiple-ring-core FMF and degenerate-mode-selective reception.

Multiple-input-multiple-output digital signal processing (DSP) has become a severe bottleneck for mode division multiplexing (MDM) because of its huge computational complexity. In this paper, we propose a novel scheme for real-time DSP-free intensity-modulation/direct-detection (IM/DD) MDM transmission, in which the transmission few-mode fiber (FMF) is characterized by multiple-ring-core structure to suppress modal crosstalk among each LP mode, while each pair of non-circularly-symmetric degenerate modes is simultaneously demultiplexed by a degenerate-mode-selective fiber coupler for DSP-free reception. Based on a 10 km ultralow-modal-crosstalk double-ring-core FMF and a pair of all-fiber 4-LP-mode MUX/DEMUX, we demonstrate the first IM/DD MDM prototype system using commercial single-mode (SM) 10 Gbps SFP + modules and 4K video transceivers without any hardware modifications. The temperature and wavelength dependence are evaluated. The stable Q2-factor performance proves that it can be a smooth evolution scheme from conventional SM IM/DD systems. Moreover, the scheme can be further extended to support more modes with improved FMF design adopting more ring areas.

Dirty-data-based alarm prediction in self-optimizing large-scale optical networks.

Machine-learning-based solutions are showing promising results for several critical issues in large-scale optical networks. Alarm (caused by failure, disaster, etc.) prediction is an important use-case, where machine learning can assist in predicting events, ahead of time. Accurate prediction enables network administrators to undertake preventive measures. For such alarm prediction applications, high-quality data sets for training and testing are crucial. However, the collected performance and alarm data from large-scale optical networks are often dirty, i.e., these data are incomplete, inconsistent, and lack certain behaviors or trends. Such data are likely to contain several errors, when collected from old-fashioned optical equipment, in particular. Even after appropriate data preprocessing, feature distribution can be extremely unbalanced, limiting the performance of machine learning algorithms. This paper demonstrates a Dirty-data-based Alarm Prediction (DAP) method for Self-Optimizing Optical Networks (SOONs). Experimental results on a commercial large-scale field topology with 274 nodes and 487 links demonstrate that the proposed DAP method can achieve high accuracy for different types of alarms.