Program

The SBDS curriculum and overall training plan are designed to prepare future leaders in multidisciplinary research that combines quantitative experimental measurements and manipulations, computational modeling, and analytical data mining.  The one-year curriculum is designed for second-year graduate students who have completed the core coursework in their degree-granting program.  Rising third-year students will also be considered eligible if they were not supported on another training grant in their second year. 

The individual components of the SBDS training plan have been spaced out to provide a consistent development experience over the period of active traineeship.  The scope is based on the expectation that SBDS trainees will dedicate 10% effort (~235 hours) to activities during their year of support.  This expectation, along with several others related to SBDS commitments, are spelled out in itemized contracts that must be signed by the primary mentor and co-mentor before trainee appointment.

All trainees will complete an 80-hour immersion in the lab of a co-mentor within the first two months of appointment to the training grant (and before the Fall semester begins). The 10 days of immersion can be split as the trainee and co-mentor see fit; for example, two full-weeks or four half-weeks. During these days, the trainee is fully embedded in the lab of the co-mentor working on the aspect of a collaborative project that directly draws in the expertise of the co-mentor’s lab. The trainee attends lab meetings and other activities of the co-mentor, physically locating themselves with the other group. The immersion is a critical feature of the SBDS training because it emphasizes the value of cross-disciplinary training, not just of technical expertise but of different lab cultures and “approaches” to research.

While preparing their application to the SBDS training program, students will identify a co-mentor with complementary scientific expertise. These mentor/co-mentor pairings will often involve the coupling of a faculty member with expertise in a particular modeling approach and a faculty member with scientific experience for a particular biomedical problem that makes use of specific modeling and data science methods.  Mentor/co-mentor pairings could also include faculty with expertise in one modeling method that is newly applied to another modeling domain. 

The mentor/co-mentor structure also allows for the ready integration of junior mentors that have not yet developed experience with graduate student training.  When the primary mentor of a trainee lacks a track record of graduated Ph.D. student trainees and secured external funding, we require that the co-mentor be a more senior, SBDS-approved faculty mentor that fills these voids. The junior-senior mentor requirement will ensure that the SBDS trainee has sufficient support to navigate their research and career development goals. An SBDS applicant with a primary mentor that is a full SBDS-approved mentor can propose any co-mentor on the approved roster. If a potential SBDS applicant would like to propose a co-mentor that is not yet an approved SBDS mentor, we will require the faculty member to be considered by the executive committee and approved as an SBDS mentor.

Since 2010, the UVA Systems Biology Journal Club has brought together students, postdocs, staff, and faculty across Grounds with a shared interest in computational and systems approaches.  It is entirely student organized and administered but mentor attended, adopting a flexible format that is geared to be maximally beneficial to the presenter.  In their year of support, SBDS students will be required to attend the Systems Biology Journal Club and present at-least once.  Trainees who have pre-completed some of the required coursework will schedule to present in the Fall semester to spread out participation.  After the support year, SBDS alumni will be expected to stay connected to the training grant through the Systems Biology Journal Club.

The ability to work collaboratively is central to the development of an SBDS cohort and a key feature of research in computational biology, bioinformatics, and data science. The SBDS training program will focus on these topics through a scheduled series of Collaborative Foundations Lunches in the Fall and Winter:

i) Principles of Collaboration in SBDS (led by Jason Papin)

ii) Collaborative Programming in SBDS (led by Nathan Sheffield)

iii) SBDS Visualization & Figure Design (led by Kevin Janes)

iv) SBDS Presentations (led by Silvia Blemker)

v) Preparation of SBDS Proposals (led by Shayn Peirce-Cottler)

The Collaborative Foundations Lunches help to establish a rapport among the trainee cohort and the larger SBDS community that helps to moderate them.

Hackathons are exceptionally powerful experiences to reinforce skills critical for successful research in the field of SBDS.  A key question is defined and cross-disciplinary student groups work together on addressing the given question.  Preference is given toward challenges that have large community engagement (e.g., DREAM challenges) and facilitate the presentation of hackathon outcomes in classes, the Systems Biology Journal Club, and other program activities with a strong presence of first-year graduate students.  The Hackathon acts as a concluding touchstone of research activity that enables SBDS trainees to recognize the skills they have gained in SBDS over the course of the training program.

The field of SBDS has many analytical approaches, posing a challenge for trainees (and mentors) to select the most-suitable method and tailor it to their particular application.  Making an informed choice requires a fundamental understanding of strengths and weaknesses based on the quantitative underpinnings of each approach.  Such knowledge is difficult for graduate students to “pick up” outside of a class that has a formal didactic component. 

Trainees benefit from a focused instruction of their own choosing, which gives them the freedom to select courses that introduce methods not ordinarily applied to their immediate research area. Together, the SBDS Foundations Selectives cover Fall and Spring semesters frequently enough that trainees can pre-complete the requirement before joining the program or enroll during their year of support.  Trainees are, of course, welcome to take a Foundations class that aligns with their research project, but they will also be encouraged to use the selective to add an orthogonal skill set.

SBDS Foundations Selectives

Option #1:  BME 6315 Systems Bioengineering (SBDS Instructors:  Papin, Peirce-Cottler, Saucerman, Fallahi-Sichani; Timing:  Spring [annual]).  This course introduces the fundamentals of systems modeling across spatial scales.  Modules focus on agent-based modeling for multicellular simulations, ordinary differential equations modeling for signal-transduction networks, flux balance analysis for metabolic models, and statistical methods for model selection.  By learning the fundamentals of several different modeling approaches paired to applications, SBDS trainees will be poised to think creatively about how such tools could be applied to their own research.
 
Option #2:  BME 6550 Molecular Data Science (SBDS Instructor:  Naegle; Timing:  Fall [annual]).  This “flipped” course integrates data wrangling and analysis from the standpoint of modern systems-biology measurements of transcriptomes, metabolomes, and proteomes.  Topics include false discovery, multifactorial designs, computational statistics, dimensionality reduction, clustering, classification, network theory, and network inference.  SBDS trainees opting for this selective will gain foundational understanding of how large, primary datasets can be used to generate testable hypotheses at the systems level.
 
Option #3:  BME 6550 Data Mining & Machine Learning for BME (SBDS Instructor:  Zhang; Timing:  Spring [odd years]).  This graduate-level course introduces the fundamental approaches, algorithms, and models of data mining and machine learning in the context of biomedical applications.  Established and leading techniques are presented and illustrated in the Python computing environment through primary biomedical data from The Cancer Genome Atlas.  Topics include feature selection, dimensionality reduction, clustering, classification, anomaly detection, transfer learning, meta-learning, and explainable AI.  Trainees who choose this fast-paced selective will gain a deeper foundation in algorithms and implementations that are established and developing within the SBDS field.
 
Option #4:  BIOC 8145 Bioinformatics & Functional Analysis of Genomes (SBDS Instructors:  Bekiranov, Zang, Ratan, Sheffield; Timing:  Spring [even years]).  This graduate-level course provides the statistical-programming background and software tools for analyzing functional genomics datasets.  The material focuses on identifying single-nucleotide and structural variants by genomics, gene-expression changes by RNA-seq and PRO-seq, and transcription-factor binding, histone modifications, and chromatin states by ChIP-seq and ATAC-seq.  The tutorials, workflows, and programming assignments are implemented in UNIX and R.  Each functional-genomics category is taught by expert SBDS mentors, ensuring transmission of the most up-to-date approaches.  This selective will give SBDS trainees the practical foundations for research projects with a heavy emphasis on genomics, transcriptomics, or epigenomics.

Required Course—BME 7370 Quantitative Biological Reasoning (SBDS Instructor: Janes; Timing: Spring [annual])

Among the most important operational skills to gain early in one’s predoctoral training is to understand how to think and learn as a graduate student.  The generality of such graduate-level operations can be reinforced through primary literature that goes beyond a single discipline, but it is critical not to sacrifice depth for breadth.  The SBDS experience will be enhanced by instilling trainees with a sense that, as a professional skill, “master of all trades” is a daunting-but-still-worthy lifelong pursuit.  Specifically, there are core concepts of experimental design, computational analysis, and results interpretation that are important and independent of a specific biomedical domain.  Detailed discussion of such concepts in various applications illustrates how similar approaches can be redeployed in very different settings.  This perspective is hugely important for SBDS, a field characterized by the rapid evolution of tools that must be critically assessed before reuse.

This course will provide students with a quantitative framework for identifying and addressing important biological questions at the molecular, cell, and tissue levels.  The course will focus on the interplay between methods and logic, with an emphasis on the themes that emerge repeatedly in quantitative experiments.  Discussions will be preceded by primary literature that illustrate how in-depth understanding of such themes led to significant conceptual advances in biochemistry, molecular biology, and cell biology.

Required course—BMS 7100 Responsible Conduct for Research (Instructor:  Eby + SBDS co-mentor discussants; Timing: Spring [annual])

The UVA School of Medicine fulfills the Responsible Conduct for Research education mandate for all of its NIH-sponsored training grants through a one-credit course offered in the Spring.  Led by the Director of Humans Subjects Research Education, the course is required for all students in their first year of training-grant support and every three years thereafter.  The format combines introductory lectures on the highest standards of practice with in-class breakout discussions that work through specific case studies relevant for the week’s topic: data ethics, human-subjects research, animal welfare, etc.  Student attendance is strictly documented, and training-grant mentors are expected to volunteer as small-group discussants for at-least one of the topics covered.