Project in Archaeology
DECODING THE SECRETS OF ANCIENT ECUADORIAN POTTERY ENGINEERS
Lue Fang, Sareena Kalinani, Saketh Karri, Lukas Liakhovitch, Amy Lin, Kaixiang Loke, Ardiv Mirza, Karthik Vemparala, Sebastian Villa, Dylan Wigdahl
Advisor: Maria Masucci, PhD
Assistant: Victoria Kuenzel
ABSTRACT
The creation of white-on-red pottery has long been attributed to the people of Guangala due to the abundance of these ceramics in the Guangala region; however, there are numerous inconsistencies that challenge this assumption. To address these discrepancies, a deeper examination of the pottery’s material composition and manufacturing method is required. This study utilizes experimental archaeology to investigate the white-on-red pottery’s forming techniques and tests different organic tempers like corn husks and balsa wood. While this unique pottery has been assumed to have originated from Guangala, Instrumental Neutron Activation Analysis (INAA) and petrographic analysis reveal non-local material compositions, suggesting alternative explanations of its presence, like trade networks with Northwest Peru or other neighboring regions. To support this suggestion, thick-section imaging identifies the pottery as highly micaceous, which is extremely uncommon for clay native to Guangala. By replicating ancient pottery techniques and comparing void size and orientation, this research demonstrates that slabbing and drawing methods, coupled with corn husk temper, achieve the characteristic void orientations and sizes found in white-on-red pottery. The research contributes to understanding the captivating white-on-red pottery and new insights into the Guangala people and their interactions with neighboring regions.
Project in Psychology
ANALYZING THE RELATIONSHIP BETWEEN INTELLECTUAL APTITUDE AND SUSCEPTIBILITY TO COGNITIVE ILLUSIONS
Ryan Buschman, Joanna Chen, Samarth Desai, Alexa Hernandez, Elbert Ho, Yeji Kim, Julia Langley, Erin Li, Grant Rupinski, Holly Sajdak, Lindsey Samuel, Rishi Venkatesh
Advisor: Patrick Dolan
Assistant: Grace Rinehart
ABSTRACT
Predictive processing allows our brains to interpret various sensory inputs through the use of assumptions rooted in prior knowledge and experiences; however, this model of processing sometimes leads to errors, known as cognitive illusions. Cognitive illusions are divided into four main categories: sensation, perception, and attention; memory; judgment and decision-making; and self-perception. To analyze their effect, we surveyed 46 online participants through CloudResearch and 48 Governor’s School of New Jersey in the Sciences (GSNJS) scholars. The survey contained a series of 16 cognitive illusions, as well as multiple personality questions to disguise the survey as a personality test (see Appendix A). Through this survey, we sought to answer the following question: given the aptitude, preparation, and motivation of GSNJS scholars, are they less susceptible to cognitive illusions than the general population? Out of the 16 analyzed illusions, GSNJS scholars were mostly equally susceptible to sensory tasks (three of five equal, one less susceptible, one more susceptible), equally susceptible to memory tasks (three of three equal), less susceptible to judgment tasks (three of six less susceptible, three equally susceptible), and more susceptible to self-perception tasks (two of two). These results suggest that GSNJS scholars demonstrate stronger academic aptitude and better decision-making skills than the general population; however, they are equally susceptible to the majority of the cognitive illusions.
T12 Final Paper >
Project in Chemistry
COMPARISON OF TRADITIONAL AND MACHINE LEARNING PROGRAMS IN THE EVALUATION OF PROTEIN-LIGAND BINDING
Blessing Anyangwe, Arushi Desai, Elizabeth Fishman, Kevin Jin, Kai Kim, Erin Kraus, Eugene Lee, Angelina Li, Bridget Liu, Nicholas Sardy, Aarna Tekriwal, Osariemen Unuigbe, Alexander Zatuchney, Eric Zhu
Advisor: David Cincotta
Assistants: Katerina Pouathas, Joel Moses
ABSTRACT
Molecular docking is an in-silico method that predicts the conformation of an interaction between two or more molecules—generally, an interaction between a ligand and its target protein. In recent years, molecular docking has had widespread applications in pharmaceutical fields, aiding in the discovery and development of drugs. Traditionally, molecular docking is performed using a search and score algorithm: the search algorithm generates a variety of protein-ligand poses, and the scoring algorithm calculates the binding strength of each pose. These results can then be used to determine the optimal binding conformation of a protein-ligand complex. More recently, molecular docking programs that rely on artificial intelligence (AI) have been developed. These deep-learning-based models—based on reverse diffusion—are trained with datasets of protein-ligand complexes. By analyzing existing protein-ligand complexes, these deep-learning molecular docking programs progressively become more successful at docking ligands to their target proteins given their molecular properties. To compare the accuracy of traditional and deep-learning-based molecular docking programs in site-specific docking (i.e. “pocket docking”), a traditional program and a deep-learning-based program were tasked with docking a set of ligands to their target proteins—all of which were not included in the deep-learning-based program’s training datasets. The traditional docking program of choice was SeeSAR “Midas” (v.13.1.1) by BioSolveIT; the deep-learning-based docking program of choice was DiffDock-Pocket, a pocket-docking version of the standard blind-docking DiffDock program. After using both programs to predict the protein-ligand binding conformations, the Root Mean Square Deviation (RMSD) value of each predicted protein-ligand conformation relative to the actual protein-ligand conformation was determined using a Python program. The RMSD values generated by the two docking programs (see Appendix B)—used to quantify docking accuracy—were then compared for all tested protein-ligand complexes. After comparing the docking accuracies of the two programs, it was determined that SeeSAR “Midas” and DiffDock-Pocket predict optimal protein-ligand binding conformations with relatively similar accuracies. Especially given the novelty of AI-based approaches in molecular docking, these methods are especially promising, and it is possible that they may surpass traditional docking programs in the near future.
T18 Final Paper >
Project in Environmental Biology
EVALUATING THE ROLE OF ZOOPLANKTON AND AQUATIC MACROINVERTEBRATES IN THE TRANSMISSION OF BATRACHOCHYTRIUM DENDROBATIDIS IN AN AMPHIBIAN DISEASE SYSTEM
Isha Amin, Maria Bonilla, Emily Hickey, Sydney Kim, Prisha Malik, Tharun Nathan,
Jessica Pappas, Rahil Patel, Sania Patel, Gabriela Ramirez, Shalini Shankar, Annette Tran
Advisor: Jessica McQuigg, Ph.D
Assistants: Harris Naqvi B.S., Jonah Fine
ABSTRACT
Discovered in 1998, the chytrid fungus Batrachochytrium dendrobatidis (Bd) has immensely decreased amphibian populations in the Americas, Europe, and Africa. Known for causing chytridiomycosis, the pathogen attacks the keratin on the skin of amphibians and inhibits their abilities to thermoregulate and remain hydrated. This study investigates the role of various zooplankton and aquatic macroinvertebrates in the transmission of Bd, and questions how, if at all, invertebrates affect the abundance of Bd in a closed system. Callibaetis ferrugineus (mayflies), Chironomidae chironomus (non-biting midges), Ferrissia fragilis (freshwater limpets), Dytiscidae (diving beetles), Cyclopoid copepod (copepods), and Hydrachnidae (water mites) were collected from Drew University’s Bd-negative artificial ponds for experimentation. Each animal was then housed in either Bd spiked or Bd negative water for the duration of the experiment. Concentrations of zoospores, the motile and infectious form of Bd, were quantified using quantitative polymerase chain reaction (qPCR) after the introduction of invertebrates to the system. Control and experimental setups were prepared for each species to assess their impact on Bd zoospore abundance over a specific period. The zoospore expression in the controls averaged ~71. The average zoospore expression for Ferrissia fragilis was significantly greater with an average Bd expression of ~382. For Callibaetis ferrugineus, the average expression was ~236. Cyclopoid copepods significantly decreased Bd presence with an average expression of ~38. These findings suggest complex interactions between invertebrates and Bd, highlighting potential biotic factors influencing pathogen dynamics in aquatic ecosystems. Further research should explore the mechanistic basis of these interactions and evaluate the potential of invertebrates in controlling Bd spread and abundance.
T20 Final Paper >
Project in Mathematical Physics
CELESTIAL MECHANICS: APPLICATION OF KEPLER’S LAWS AND SPHERICAL TRIGONOMETRY
William Colangelo III, Adrien Cristian, Stefano D’Agostino, Mayank Deoras, Rishay Gupta, Tyler Harms, Jeffrey Jiang, David Ji, Aditya Kirubakaran, Krish Shah, Timothy Torubarov, Kevin Zhang
Advisor: Steve Surace
Assistant: Clifford Wijaya
ABSTRACT
Astronomy has evolved from qualitative observations to a precise science, founded on the basic laws of gravitation. Isaac Newton’s law of universal gravitation and Kepler’s laws of planetary motion remain vital for predicting celestial movements. This paper applies these principles to calculating the positions of Earth and and other celestial bodies on any specific date, the exact time of sunset and sunrise, and predicting the times at which celestial bodies become visible each day. By deriving equations using Newton’s and Kepler’s laws and exploring geometric represen- tations of elliptical orbits, the relevance of Newtonian mechanics in contemporary astronomy is highlighted. This study provides practical insights into the motion of planets and a very close ap-proximation of the orbits of celestial objects in the solar system.