2024 CompX Recipients
About the CompX Faculty Grants Program
Winners of the 2024-2025 Neukom Institute CompX Faculty Grants Program for Dartmouth faculty have been announced for one-year projects. We received over $1.2million in total requests and awarded a total of $250K.
The program seeks to fund both the development of novel computational techniques as well as the application of computational methods to research across the campus and professional schools.
Dartmouth College faculty including the undergraduate, graduate, and professional schools were eligible to apply for these competitive grants.
Note: * indicates an award that is partnered with assistance from Dartmouth College Research Computing.
+ indicates an award that is partnered with an RA provided through the Scholars program
Anthropology: Raquel Fleskes & Reed Harder
Beyond Bones: Non-Destructive Insights into Population History through Ancient DNA Analysis of Clay Tobacco Pipestems
Advances in the extraction and analysis of ancient DNA (aDNA) from human remains has provided a powerful tool to understand human population history. However, protocols for obtaining aDNA typically require bone or teeth samples from human remains, which are regarded as materials of cultural sensitivity and are in limited quantity. Alternative materials, such as DNA extracted from archaeological objects, present a means of investigating genomic population history without the use of human remains. Clay tobacco pipestems are common artifacts in colonial period archaeological sites in North America, and have been shown to preserve the DNA of people who smoked from them. Methods for reliably extracting and analyzing aDNA in these pipestems could significantly expand the ability of researchers to study colonial population history while avoiding issues of destructive sampling of human skeletal remains.
With this project, our aim is to create an optimized protocol for aDNA extraction from clay tobacco pipestems which encompasses both laboratory and bioinformatic methods. This aim will be accomplished via deeper sequencing of already generated pipestem DNA extracts with additional targeted human genomic capture. The newly generated data, in combination with previously generated DNA sequencing reads, will be analyzed to investigate population history, settlement patterns, interactions and health of colonial-era persons at North American archaeological sites.
Biological Sciences: Robert Hill & Aman Aberra
Determining and Enhancing the Cellular Selectivity of Transcranial Magnetic Stimulation
Transcranial magnetic stimulation (TMS) is a safe, noninvasive technique to modulate brain activity in humans. TMS works by generating a magnetic field with an electromagnetic coil placed outside the head, which induces a pulsed electric field (E-field) in the brain. TMS-based therapies are FDA-approved for drug-resistant depression, obsessive compulsive disorder, and migraine. However, responses to TMS exhibit high variability and low effect sizes, meaning many patients fail to experience relief. Current approaches to selecting TMS dose assume uniform activation of different neuronal subtypes, despite considerable diversity in their morphologies, electrical properties, and functional roles. It remains unclear how to select the optimal E-field dose to target specific neural circuits due to gaps in our understanding of how the E-field alters brain activity of diverse neuronal subtypes.
To address this gap, this project will use optical techniques for resolving neural activity and subcellular structure during stimulation with precisely controlled E-fields. Using this approach, we will measure the activation of different neuronal subtypes as a function of E-field pulse waveform and intensity. In addition, previous experimental and modeling work suggests neurons with axonal myelin, an insulating layer of fatty membrane, are more sensitive to TMS. Therefore, we will also use a label-free technique to visualize myelin, allowing us to determine how neuronal subtype and myelination correlate with the responses to E-field stimulation. Finally, these functional and structural data will be used to build biophysically-realistic computational models capturing the E-field dose–response relationship of different neurons with the goal of identifying pulse waveforms that could achieve more selective activation of specific neuron populations.These results will provide fundamental insights into the response of individual neurons to electric fields, which is a critical step to designing more effective TMS stimulation protocols informed by the underlying neural mechanisms.
Linguistics: Christiane Donahue
A Database of Student Writing for Longitudinal Writing Research
Research in the US field of Writing Studies focused on student writing development has been, by and large, notoriously resistant to computational techniques until recent years--even as corpus linguistics has long demonstrated the value of these techniques to writing research. At Dartmouth, we are undertaking writing research using mixed methods to carefully analyze change over time in student writing. Students have the opportunity to create digital WordPress portfolios in which they can choose to feature their work over their time at Dartmouth. The portfolio is highly valuable to the students, while also enabling (with their consent) a corpus of student writing that serves as a source for multiple research projects about student writing growth, including projects using computational techniques.
Classics: Julie Hruby
Associating Fingerprint Patterns with Age and Sex: A Quantifiable Approach
A wide range of prehistoric and ancient Greek ceramic objects, including vessels, ceramic sculpture, seal impressions, and writing tablets preserve the fingerprint impressions of their producers. Traditionally, archaeologists have matched prints in order to understand ancient labor systems, but more recently, we have also begun to ask a much wider range of questions. The ages and sexes of producers are among those questions, but so far, the techniques that have been used to reconstruct those factors have typically been able to work on the level of populations rather than individuals, and they have also been subject to challenges posed by differential clay shrinkage rates (among other things).
The current project will improve the accuracy of sexing and aging producers of ancient Greek artifacts by using fingerprints that were accidentally impressed in objects made by modern Greek ceramicists as a reference sample. Fingerprint impressions from modern Greek adult potters of known sexes and age grades have already been collected and scanned with a high-resolution 3D scanner, cropped, and exported. A Greek attorney has assisted us in complying with both European Union and Greek law as they relate to the collection of prints from juveniles. We are continuing the process of collecting and scanning modern prints from juveniles. Our CompX grant will also facilitate renewal of fingerprint matching software.
Geography: Justin Mankin & Maggie O'Shea
Data Poverty in Climate Damage Assessments
Even in this era of big data, data poverty in weather and climate measures abound. Maggie O'Shea (EEES PhD student) and Justin Mankin (GEOG) combine weather, climate, and socioeconomic data in a simulation framework to diagnose the costs of data gaps for detecting, attributing, pricing, and managing climate impacts.
Geisel School of Medicine: Shrey Sukhadia & Kristen Muller
Transformative Pathomics in Oncology: Harnessing Pathology Tissue Slide Imagery for Clinically Relevant Gene Expression Prediction in Invasive Breast Carcinoma
Pathologists analyze H&E-stained Formalin fixed paraffin embedded (FFPE) tissue slides under a microscope routinely to study the morphological patterns of tumor tissue and surrounding stroma in order to arrive at the prognoses or diagnoses for cancer patients. Pathomics is the study of extracting quantitative imaging features such as morphometric, intensity and gradient-based features from such tissue slides and holds a promise to elevate prognostic and diagnostic accuracy by revealing intricate tissue and cellular details by having a better (quantitative) view of the tissue-characteristics. Our research aims to advance the concept of pathomics by evaluating how pathomic features extracted from tissue imagery could predict gene-expression derived from sequencing of RNA material extracted from the respective tissues embedded in the tissue-slides. Such a prediction could be performed using machine learning that includes both supervised and non-supervised modeling techniques.
Such techniques may open up avenues for the discovery of novel genes (or, biomarkers) and their contribution towards the tissue-characteristics. Studying the role of these biomarkers in biological pathways necessary for the function of cells may provide a direct correlation of tissue-characteristics with cellular-mechanisms relevant to cancer. This would also offer advancements in the molecular profiling and targeted therapy in Invasive Breast Carcinoma.
Biology Sciences and Geography: Sophie von Fromm
Changes in Soil Organic Carbon Dynamics – insights from modeling 14C timeseries at Hubbard Brook Experimental Forest
Soil is one of the most important carbon (C) reservoirs on planet Earth. Each year, ~30% of CO2 emitted to the atmosphere is sequestered by terrestrial ecosystems, where its majority becomes soil organic carbon (SOC). To study the soil C cycle and its response to global change, model development is needed to accurately represent the amount and timescales of C transferred between different reservoirs. The timescales over which SOC responds to changes in climate and vegetation remain one of the largest uncertainties in the terrestrial C cycle. To overcome this uncertainty, long-term timeseries of soil C isotopic (14C) data are needed to accurately model soil C stocks and C turnover rates, allowing for a better understanding of changes in soil C dynamics over time and under global change.
In this study, we will use archived soil samples from the Hubbard Brook Experimental Forest (HBEF) from 1998 to the present. The long-term HBEF site is ideal for developing a data-driven model that can accurately represent the magnitude of C transfer between different soil pools and the atmosphere over time. Radiocarbon, which can be used as a tracer of the soil C cycle, will be measured on 135 soil samples from different depth layers and years to constrain compartment-based SOC decomposition models. This will allow us to i) accurately model soil C turnover rates from radiocarbon timeseries data, and ii) test whether changes in C turnover rates at HBEF are related to the observed environmental changes. This project will allow us to identify the most appropriate model structure to represent non-steady soil C dynamics at HBEF. Very few places on Earth can provide such sample material and experimental set-up. Our improved understanding of soil C dynamics will enable us to assess the potential of northern hardwood forest soils to sequester C from the atmosphere as a nature-based solution to mitigate climate change.
Psychological & Brain Sciences: Peter Tse & Jay Vincelli
Octopus Electroencephalography
Our goal is to invent underwater electroencephalography (EEG) for measuring the brainwaves of octopuses in order to measure correlates of octopus neural activity non-invasively. We are able to detect stimulus frequency-dependent correlates of flickering LEDs in electrodes placed on the skin located between the eyes of octopus bimaculoides. This has proven to be challenging because of the inability of classical EEG techniques to measure reliable signals underwater due to the high electrical conductance of seawater and EEG's exquisite sensitivity to common-mode sources of noise such as muscle artifacts that are not easily controlled for in active, slippery animals such as octopuses.
In addition, octopuses lack bones, making fixed electrodes impractical. To overcome this, rather than placing electrodes onto the skin of an octopus, we place the octopus between two layers of fixed electrodes underwater, while presenting the octopus with visual input from outside its enclosure. Because the central brain is located just under the skin between the two eyes, we can measure brainwaves from this location. Indeed, we can measure frequency tagged responses to ongoing stimuli in a manner that has not been attempted before in the octopus. This CompX award will be used to support a post-doc engineer, Jay Vincelli, for one year of EEG data collection and data analysis using the breakthrough technology developed by Jay and collaborators at the University of Rhode Island School of Engineering and the University of Nevada at Reno.
Government: Sean Westwood
Understanding What Elected Officials Say
Partisan animosity is on the rise in the United States, potentially threatening democracy and the ability of Americans to engage with those who hold different political views. This problem has, thus far, been studied at the national level. This project, by creating the most comprehensive database on what state-level elected officials say and do while in office, will enable deep exploration of partisan animosity in the American states.
This CompX grant enables us to hire and train Dartmouth undergraduates to collect and analyze data from all state-level elected officials. Research assistants will scour sources of rhetoric data from elected officials in all 50 states. They will work with us to deploy LLMs to categorize the text in real time. Our LLMs are fine tuned to identify the specific content and context of each text (e.g., allowing us to distinguish between opposition to policy and personal attacks). With these data will identify both lawmakers who stoke conflict and those who are constructive policy discourse.
The long-term outcomes of this project are academic work, open data, and a public-facing dashboard that displays data on all 7,000+ elected state officials across 50 states. Users will be able to see those representatives who are spreading divisiveness versus those focusing on the business of government.