Electrical Engineering and Computer Science Scholarly Workshttps://hdl.handle.net/1808/3852024-03-29T07:06:16Z2024-03-29T07:06:16ZEditorial: Non-coding RNAs: insights and state-of-the-art in gastrointestinal sciencesFu, TingXu, Zhenjiang ZechZhong, Cunconghttps://hdl.handle.net/1808/347362023-08-16T06:06:54Z2023-07-05T00:00:00ZEditorial: Non-coding RNAs: insights and state-of-the-art in gastrointestinal sciences
Fu, Ting; Xu, Zhenjiang Zech; Zhong, Cuncong
2023-07-05T00:00:00ZGender, Smoking History, and Age Prediction from Laryngeal ImagesZhang, TianxiaoBur, Andrés M.Kraft, ShannonKavookjian, HannahRenslo, BryanChen, XiangyuLuo, BoWang, Guanghuihttps://hdl.handle.net/1808/345792023-07-12T06:06:54Z2023-05-29T00:00:00ZGender, Smoking History, and Age Prediction from Laryngeal Images
Zhang, Tianxiao; Bur, Andrés M.; Kraft, Shannon; Kavookjian, Hannah; Renslo, Bryan; Chen, Xiangyu; Luo, Bo; Wang, Guanghui
Flexible laryngoscopy is commonly performed by otolaryngologists to detect laryngeal diseases and to recognize potentially malignant lesions. Recently, researchers have introduced machine learning techniques to facilitate automated diagnosis using laryngeal images and achieved promising results. The diagnostic performance can be improved when patients’ demographic information is incorporated into models. However, the manual entry of patient data is time-consuming for clinicians. In this study, we made the first endeavor to employ deep learning models to predict patient demographic information to improve the detector model’s performance. The overall accuracy for gender, smoking history, and age was 85.5%, 65.2%, and 75.9%, respectively. We also created a new laryngoscopic image set for the machine learning study and benchmarked the performance of eight classical deep learning models based on CNNs and Transformers. The results can be integrated into current learning models to improve their performance by incorporating the patient’s demographic information.
2023-05-29T00:00:00ZOptical trapping of sub-millimeter sized particles and microorganismsLialys, LaurynasLialys, JustinasSalandrino, AlessandroAckley, Brian D.Fardad, Shimahttps://hdl.handle.net/1808/343272023-06-13T06:06:33Z2023-05-27T00:00:00ZOptical trapping of sub-millimeter sized particles and microorganisms
Lialys, Laurynas; Lialys, Justinas; Salandrino, Alessandro; Ackley, Brian D.; Fardad, Shima
While optical tweezers (OT) are mostly used for confining smaller size particles, the counter-propagating (CP) dual-beam traps have been a versatile method for confining both small and larger size particles including biological specimen. However, CP traps are complex sensitive systems, requiring tedious alignment to achieve perfect symmetry with rather low trapping stiffness values compared to OT. Moreover, due to their relatively weak forces, CP traps are limited in the size of particles they can confine which is about 100 μm. In this paper, a new class of counter-propagating optical tweezers with a broken symmetry is discussed and experimentally demonstrated to trap and manipulate larger than 100 μm particles inside liquid media. Our technique exploits a single Gaussian beam folding back on itself in an asymmetrical fashion forming a CP trap capable of confining small and significantly larger particles (up to 250 μm in diameter) based on optical forces only. Such optical trapping of large-size specimen to the best of our knowledge has not been demonstrated before. The broken symmetry of the trap combined with the retro-reflection of the beam has not only significantly simplified the alignment of the system, but also made it robust to slight misalignments and enhances the trapping stiffness as shown later. Moreover, our proposed trapping method is quite versatile as it allows for trapping and translating of a wide variety of particle sizes and shapes, ranging from one micron up to a few hundred of microns including microorganisms, using very low laser powers and numerical aperture optics. This in turn, permits the integration of a wide range of spectroscopy techniques for imaging and studying the optically trapped specimen. As an example, we will demonstrate how this novel technique enables simultaneous 3D trapping and light-sheet microscopy of C. elegans worms with up to 450 µm length.
2023-05-27T00:00:00ZIntegrated de novo gene prediction and peptide assembly of metagenomic sequencing dataThippabhotla, SirishaLiu, BenPodgorny, AdamYooseph, ShibuYang, YoungikZhang, JunZhong, Cunconghttps://hdl.handle.net/1808/340832023-04-11T06:07:04Z2023-03-11T00:00:00ZIntegrated de novo gene prediction and peptide assembly of metagenomic sequencing data
Thippabhotla, Sirisha; Liu, Ben; Podgorny, Adam; Yooseph, Shibu; Yang, Youngik; Zhang, Jun; Zhong, Cuncong
Metagenomics is the study of all genomic content contained in given microbial communities. Metagenomic functional analysis aims to quantify protein families and reconstruct metabolic pathways from the metagenome. It plays a central role in understanding the interaction between the microbial community and its host or environment. De novo functional analysis, which allows the discovery of novel protein families, remains challenging for high-complexity communities. There are currently three main approaches for recovering novel genes or proteins: de novo nucleotide assembly, gene calling and peptide assembly. Unfortunately, their information dependency has been overlooked, and each has been formulated as an independent problem. In this work, we develop a sophisticated workflow called integrated Metagenomic Protein Predictor (iMPP), which leverages the information dependencies for better de novo functional analysis. iMPP contains three novel modules: a hybrid assembly graph generation module, a graph-based gene calling module, and a peptide assembly-based refinement module. iMPP significantly improved the existing gene calling sensitivity on unassembled metagenomic reads, achieving a 92–97% recall rate at a high precision level (>85%). iMPP further allowed for more sensitive and accurate peptide assembly, recovering more reference proteins and delivering more hypothetical protein sequences. The high performance of iMPP can provide a more comprehensive and unbiased view of the microbial communities under investigation. iMPP is freely available from https://github.com/Sirisha-t/iMPP.
2023-03-11T00:00:00Z