The RIP series provides an informal forum for computational biologists to keep abreast of colleagues' projects, to help students and postdocs hone their presentation skills, and to get expert feedback on new projects. The forum is targeted towards anyone working at the interface of Biology and Analytical sciences. CBBG students are strongly encouraged to present. These meetings will be held in room #4105 Brendan Iribe Center (#432) Thursdays from 3-4pm, unless otherwise noted. There will be two presentations, approximately 30 minutes each with time for open discussion and socializing. Sign up here.
**If you are interested in speaking on an open date, or if you need to cancel or reschedule, please contact
Date |
Speaker |
Fac/
Postdoc/
Grad |
PI/Lab/Host |
Topic & Abstract |
Time (if other than 3pm) |
|---|
|
|
5/23/19 |
Nicholas Rachmaninoff |
Grad |
Tsang(NIH)/Corrada Bravo Lab |
Systems immunology of monogenic immune disorders
Abstract: Many well-known, highly penetrant, single-gene mutations can lead to dysregulation of the immune system. These range from primary immunodeficiencies (PIDs) such as Chronic Granulomatous Disease (CGD; mutations in NADPH oxidase genes) and Job’s Syndrome (STAT3 dominant-negative mutations), to autoinflammatory conditions such as Familial Mediterranean Fever (FMF; MEFV mutations) with genetic defects in inflammasome regulation. We undertook a systems approach, including analysis of blood transcriptome (whole-blood microarray) and proteome (Somascan 1.3K), to evaluate approximately 250 patients with ~20 distinct monogenic disorders, spanning a spectrum of PID and autoinflammatory conditions. These diseases provide a rich phenotypic spectrum for analyzing the major sources of variation in the human immune system. Using a variance components model we find that patient-to-patient differences explain more variation in these data than differences between conditions. Additionally, utilizing a previously described method, JIVE, we found an axis of variation shared between the transcriptomic data and proteomic data that robustly separates healthy individuals from disease. We validated the prognostic value of this signature in public data consisting of 23 studies of 4 different autoimmune diseases. (This research was supported by the Intramural Research Program of NIH, NIAID)
|
Room 4105 Iribe Center |
|
5/23/19 |
Domenick Braccia |
Grad |
Corrada Bravo Lab |
Covariate based approaches for multiple hypothesis testing correction
Abstract: Before the advent of next generation sequencing (NGS) technologies, biological experiments could typically be evaluated statistically using classical hypothesis testing procedures. However, the number of hypotheses we can generate and test in the NGS era has risen to hundreds and even millions in the case of genome wide association studies. It is inevitable that with thousands of hypotheses being tested, some of those will turn out to be false positive discoveries. Approaches to correcting false discoveries have been considered for decades (e.g. Bonferroni Correction 1935-36, and Benjamini-Hochberg 1995, among others) but other than p-values generated for each hypothesis test, they largely do not consider any additional information about the system being studied. This lack of required input information has also contributed to their popularity, making them widely applicable to many different fields challenged by multiple testing correction. Recently, however, modern multiple hypothesis testing methods have been shown to increase power while still controlling global false discovery rate. Here, I will present preliminary results on applying a covariate based approach - Independent Hypothesis Weighting (Ignatiadis 2016 et al.) - to differential abundance tests of bacteria in children from developing countries with moderate to severe diarrhea. |
Room 4105 Iribe Center |
|
5/30/19 |
Cindy Li |
Grad |
Leiserson Lab |
TBA |
Room 4105 Iribe Center |
|
5/30/19 |
Muzi Li |
Grad |
Liu/Mount Lab |
TBA |
Room 4105 Iribe Center |
