A sufficient amount of data is crucial for high-performance and accurate trend prediction. However, it is difficult and time-consuming to collect agricultural data over long periods of time; the consequence of such difficulty is datasets that are characterized by missing data. In this study we use a time-series generative adversarial network (TimeGAN) to synthesize multivariate agricultural sensing data and train RNN (Recurrent Neural Network), LSTM (Long Short-Term Memory), and GRU (Gated Recurrent Unit) neural network prediction models on the original and generated data to predict future pest populations. After our experiment, the data generated using TimeGAN and the original data have the smallest EC value in the GRU model, which is 9.86. The results show that the generative model effectively synthesizes multivariate agricultural sensing data and can be used to make up for the lack of actual data. The pest prediction model trained on synthetic data using time-series data generation yields results that are similar to that of the model trained on actual data. Accurate prediction of pest populations would represent a breakthrough in allowing for accurate and timely pest control.

peer-reviewed) is the author/funder. All rights reserved. No reuse allowed without permission.The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/154880 doi: bioRxiv preprint first posted online 3 During embryonic development, complex tissue-scale movements emerge from 49The cardiac lineage in the tunicate Ciona provides the simplest example of collective 50 cell migration [13][14][15][16] . In Ciona, multipotent cardiopharyngeal progenitors derive from the 51 B7.5 blastomeres in 110-cell embryos, and later produce fate-restricted heart and 52 pharyngeal muscle precursors [17][18][19][20] . On either side of tailbud embryos, pairs of 53 cardiopharyngeal progenitor cells (aka trunk ventral cells, TVCs) collectively polarize 54 and migrate between the ventral epidermis and the trunk endoderm, until they stop and 55 produce distinct fate-restricted progenitors 13, 14, 16, 19 . During migration, the leader TVC 56 extends dynamic protrusions and generates a broad leading edge, while the trailer 57 terminates in a tapered retraction end 13, 16 . Before migration, the surrounding trunk 63cardiopharyngeal fate specification and TVC migration 15, 16, 18, 21 , and the transcriptome of 64 migratory TVCs have been extensively profiled 16,[22][23][24] . Therefore, TVC migration provides 65 peer-reviewed) is the author/funder. All rights reserved. No reuse allowed without permission. Figure S1, Movie 5), we thus excluded Egfr from 103 further analysis ( Figure S1). Ddr, Vegfr, and Fgfr transcripts were detected in migrating 104TVCs ( Figures 1B, S1). Vegfr and Fgfr were expressed in B7.5 lineage founder cells, while 105newborn TVCs upregulated Ddr before the onset of collective migration ( Figure S1). 107The transcription factor Foxf was proposed to act as a key transcriptional regulator of 139To study the functions of selected RTKs during TVC migration, we generated passive 148These observations indicate that proper function of the Foxf target Ddr is required for 149 maintenance of collective polarity and directional migration of the TVCs. 151TVCs expressing Foxf-driven dnVegfr were also able to initiate migration, and their The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/154880 doi: bioRxiv preprint first posted online 7 only increased the standard deviation of the LT angle relative to the sagittal plane, which 160 is consistent with cells remaining constrained in at least one dimension ( Figure 1G). 162 Ddr promotes integrin-based cell-matrix adhesion to the epidermis 163We first focused on understanding how Ddr regulates TVC polarity and migration, 164and analyzed changes in cell morphology and contacts with surrounding tissues 13 . We Figure 5). We generated a dominant negative form of Intβ1 (dnIntβ1) 216To further characterize the functional relationships between the two collagen-receptors, 217we asked whether Intβ1 is required for Ddr activation. To assay localization and 218 activation of full-length Ddr, we used the minimal Foxf T...

KRAS is the most frequently mutated oncogene in human cancer, and KRAS inhibition has been a longtime therapeutic goal. Recently, inhibitors (G12C-Is) that bind KRASG12C-GDP and react with Cys-12 were developed. Using new affinity reagents to monitor KRASG12C activation and inhibitor engagement, we found that, reflecting its action upstream of SOS1/2, SHP2 inhibitors (SHP2-Is) increased KRAS-GDP occupancy, enhancing G12C-I efficacy. SHP2-Is abrogated feedback signaling by multiple RTKs and blocked adaptive resistance to G12C-Is in vitro, in xenografts, and in syngeneic KRASG12C-mutant pancreatic ductal adenocarcinoma (PDAC) and non-small cell lung cancer (NSCLC) models. Biochemical analysis revealed enhanced suppression of ERK-, MYC-, anti-apoptotic-, and cell-cycle genes, and increased pro-apoptotic gene expression in tumors from combination-treated mice. SHP2-I/G12C-I also evoked favorable changes in the immune microenvironment, decreasing myeloid suppressor cells, increasing CD8+ T cells, and sensitizing tumors to PD-1 blockade. Experiments using cells expressing inhibitor-resistant SHP2 showed that SHP2 inhibition in PDAC cells is required for tumor regression and remodeling of the immune microenvironment, but also revealed direct inhibitory effects on angiogenesis resulting in decreased tumor vascularity. Our results demonstrate that SHP2-I/G12C-I combinations confer a substantial survival benefit in PDAC and NSCLC and identify additional combination strategies for enhancing the efficacy of G12C-Is.

Purpose: To report on the design and operation of a fluoroscopic automatic gain control (AGC) system with arbitrary‐shaped operator‐selected region of interest (ROI) for a new high resolution detector consisting of an array of two EMCCD‐based modules. Methods: A solid‐state x‐ray imager consisting of dual EMCCD‐based modules was built to provide high resolution (effective pixel size 26.4 um) and wide dynamic range (55dB) in fluoroscopy. A graphic user interface was employed based on LabVIEW to display the 2K × 1K 12‐bit digital images. The image from the second sensor was digitally rotated and shifted so as to maintain image registration with the image from the first sensor. The AGC system consisting of a real‐ time feed‐back loop and a universal asynchronous receiver/transmitter platform was implemented to enable control by comparing the current average digital values in the operator‐selected ROI with a preset desired value resulting in a step by step change of the EMCCD built‐in gains to achieve the desired image brightness. A standard aluminum step wedge was used to demonstrate the response speed and a head phantom was employed to provide a realistic complex structure, rich in contrast variations. Results: The operator‐selected ROI was able to be positioned anywhere in the full 1K × 2K field of view, even with parts in each module. The settling time for smoothly reaching stable brightness was in the range of 10 to 20 frames. A video sequence demonstrates a smooth brightness transition to the desired value when the head phantom is arbitrarily moved across the field of view of the detector. Conclusions: We introduce a new real‐time AGC design for a unique custom‐built high resolution modular array detector. Stable levels of brightness within the ROI were demonstrated. Operators can set up an arbitrary shaped ROI in advance or during the procedure. Support: NIH Grants R01‐EB008425, R01‐EB002873