SCHOOL OF MECHANICAL ENGINEERING SEMINAR

Monday 4.03.2024 at 14:00

Wolfson Building of Mechanical Engineering, Room 206

Asymmetry between the dorsal and ventral digging valves of the female locust: function and mechanics

Samuel Gershon

M.Sc. Student, under the supervision of Bat El Pinchasik

School of Mechanical Engineering, Tel Aviv University, Tel Aviv, Israel

Some acridid insect, such as locusts and grasshoppers, possess an ovipositor that is highly specialized for digging and burrowing into the ground in order to lay eggs into holes in the ground. The abdomen extends into the ground during the digging while the insect remains stationary. At the end of the abdomen there are two pairs of valves that are especially adapted for preforming the burrowing of the abdomen. The valve pairs open and close cyclically in the ground in order to remove soil and extend the hole in the ground. The valve pairs, while being generally similar in shape, also possess noticeable asymmetries, which are likely related to different roles in the burrowing process. The dorsal valves are believed to be used mostly for soil shoveling and compression to the side of the dug hole; while, the ventral valve’s main function is assumed to be acting as anchor and leverage during the abdomen extension into the ground.

In this work I replicated and quantified the geometry of the valves using a mathematical model I developed. I developed different methods to automatically fit the mathematical model to 3D models I created from μCT scans. I also used finite elements analysis in order to recreate the mechanical response of the valves, and to compare the response of the ventral and dorsal valves. Using both the mathematical model and FEA I then investigated the significance of geometrical features of the valves by altering them beyond their natural shape. Specifically, I found that a geometrical feature, the skew of the valve cross sections was behaving differently between the dorsal and ventral valve. The effect of this feature was found to be negligible in terms of the load bearing capabilities, however, I found that it does effect load distribution during soil shoveling which suggests that this feature evolved to fit the function of the dorsal valve as a soil shoveling and compressing implement. The geometrical model also presents opportunities for future design and improvement of digging tools based on the locust valves but altered in a controlled manner to improve specific functionalities.

Electrical Engineering Systems Seminar

Speaker: Ilya Margolin

M.Sc. student under the supervision of Prof. Noam Shomron and Prof. Shai Avidan

Wednesday, 28^th February 2024, at 15:30

Room 011, Kitot Building, Faculty of Engineering

Enhancing Cancer Classification through cfDNA and Transformer Models

Abstract

Cancer diagnosis often faces the critical challenge of late detection and advanced metastasis, leading to complex treatments and high mortality rates. Liquid biopsy, a method for extracting cell-free DNA (cfDNA) from blood, offers a promising avenue for the early detection of cancer through the identification of biomarkers in a non-invasive and cost-effective manner. However, early detection using tumor-derived cfDNA presents formidable hurdles, primarily due to the challenge of distinguishing tumor-derived cfDNA in a normal background without complex sequencing or reference samples.

Transformer-based deep learning models, renowned for their sequence modeling capabilities, have achieved remarkable success in diverse domains. Particularly, end-to-end models demonstrate the capacity to learn significant features and exhibit enhanced robustness. In this study, an end-to-end deep learning method based on a transformer encoder was developed and applied to multiple clinical datasets containing raw genomic sequencing data from cancer patients and healthy individuals. To simulate ultra low-cost, low-coverage sequencing, the method was further tested on down-sampled sequencing samples.

For multi-cancer classification, our method achieved 79% sensitivity at 85% specificity. Using a breast cancer specific model rather than the multi-cancer model improved breast cancer classification sensitivity from 68% at 85% specificity to 86% at 85% specificity. While traditional cancer detection methods show a dramatic drop in performance with lower sequencing coverage, our method shows almost similar cancer detection ability across all tested depths.

Transformer based deep learning models show potential for direct analysis of raw genomic data. With sufficient training data, optimized models may approach accuracies needed for clinical use. Our results show the feasibility of detecting cancer at an early stage directly from cfDNA. Notably, our technique requires an extremely small amount of sequencing data. This enables the method to be scaled for widespread early cancer screening applications.

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר