White-Box Machine Learning

Recent advances in high-throughput experimental technologies and data analyses have enabled unprecedented observation, quantification and association of biological signals with cellular and clinical phenotypes. However, existing methods for extracting biological information from large biomedical datasets frequently encode such relationships in opaque "black-boxes" lacking mechanistic structure. We are developing integrated network biology, biochemical screening and machine learning approaches for revealing transparent "white-box", causal molecular mechanisms underlying drug efficacy and disease pathogenesis.

Relevant Publications

  • Yang JH, Wright SN*, Hamblin M*, McCloskey D, Alcantar MA, Schrübbers L, Lopatkin AJ, Satish S, Nili A, Palsson BO, Walker GC, Collins JJ. A white-box machine learning approach for revealing antibiotic mechanisms of action. Cell. 2019;177(6):1649-1661. [link]

Context-Dependence in Antibiotic Efficacy

The human body is a complex and dynamic ecosystem comprised of multiple species, cell types and biochemical niches. We are employing quantitative metabolomics and statistical modeling to better understand cell and metabolite dynamics at the host-pathogen interface and the effects of environmental factors on the treatment of infectious disease.

Relevant Publications

  • Yang JH*, Bhargava P*, McCloskey D, Mao N, Palsson BO, Collins JJ. Antibiotic-induced changes to the host metabolic environment inhibit drug efficacy and alter immune function. Cell Host Microbe. 2017; 22(6):757-765. [link]

  • Yang JH*, Bening SC*, Collins JJ. Antibiotic efficacy – context matters. Curr Opin Microbiol. 2017; 39:73-80. [link]

  • Meylan S, Porter CB*, Yang JH*, Belenky P, Gutierrez A, Lobritz MA, Park J, Kim S, Moskowitz S, Collins JJ. Carbon sources tune antibiotic susceptibility in Pseudomonas aeruginosa via Tricarboxylic acid cycle control. Cell Chem Biol. 2017; 24(2):195-206. [link]

Mechanisms of Antibiotic Lethality


Antimicrobial resistance is a growing threat to global health. Many studies have shown that in addition to inhibiting essential functions, antibiotics elicit death through secondary effects on cell metabolism. We are integrating the use of molecular biology, chemical screens, metabolic modeling and data-driven learning to map out the relationship between cell metabolism and antibiotic sensitivity and to identify new ways to metabolically resensitize genetically resistant bacteria to antibiotics.


Relevant Publications

  • Takahashi N*, Gruber CC*, Yang JH, Liu X, Braff D, Yashaswini C, Bhubhani S, Furuta Y, Andreescu S, Collins JJ, Walker GC. Lethality of MalE-LacZ hybrid protein shares mechanistic attributes with oxidative component of antibiotic lethality. Proc Natl Acad Sci U S A. 2017; 114(34):9164-9169. [link]

  • Lobritz MA*, Belenky P*, Porter CBM, Gutierrez A, Yang JH, Schwarz EG, Dwyer DJ, Khalil AS, Collins JJ. Antibiotic efficacy is linked to bacterial cellular respiration. Proc Natl Acad Sci U S A. 2015; 112(27):8173-80. [link]

  • Dwyer DJ*, Belenky PA*, Yang JH*, MacDonald IC, Martell JD, Takahashi N, Chan CT, Lobritz MA, Braff D, Schwarz EG, Ye JD, Pati M, Vercruysse M, Ralifo PS, Allison KR, Khalil AS, Ting AY, Walker GC, Collins JJ. Antibiotics induce redox-related physiological alterations as part of their lethality. Proc Natl Acad Sci U S A. 2014; 111(20):E2100-9. [link]

  • Lopatkin AJ, Stokes JM, Zheng EJ, Yang JH, Takahashi MK, You L, Collins JJ. Bacterial metabolic state more accurately predicts antibiotic lethality than growth rate. Nat Microbiol. In press. [link]

Context-Dependence in Cardiac Cell Signaling

All living organisms possess an incredible ability to integrate many simultaneous (and sometimes contradictory) signals to make complex decisions in selecting between different biological phenotypes. Some of the mechanisms underlying these behaviors include spatio-temporal localization and topological features of biological networks. We are integrating the use of dynamical cell modeling and live cell imaging to better understand these mechanisms on a quantitative level and to explore their effects on cardiac physiology and disease pathogenesis.


Relevant Publications

  • Yang JH, Polanowska-Grabowska RK, Smith JS, Shields CW, Saucerman JJ. PKA catalytic subunit compartmentation regulates contractile and hypertrophic responses to β-adrenergic signaling. J Mol Cell Cardiol. 2014; 66:83-93. [link]

  • Sample V*, DiPilato LM*, Yang JH*, Ni Q, Saucerman JJ, Zhang J. Regulation of nuclear PKA revealed by spatiotemporal manipulation of cAMP. Nat Chem Biol. 2012; 8(4):375-82. [link]

  • Yang JH, Saucerman JJ. Phospholemman is a negative feed-forward regulator of Ca2+ in β-adrenergic signaling, accelerating β-adrenergic inotropy. J Mol Cell Cardiol. 2012; 52(5):1048-55. [link]

  • Yang JH, Saucerman JJ. Computational models reduce complexity and accelerate insight into cardiac signaling networks. Circ Res. 2011; 108(01):85-97. [link]

  • Benedict KF, Mac Gabhann F*, Amanfu RK*, Chavali AK*, Gianchandani EP*, Glaw LS*, Oberhardt MA*, Thorne BC*, Yang JH*, Papin JA, Peirce SM, Saucerman JJ, Skalak TC. Systems analysis of bounded signaling modules generates experimental roadmap for eight major diseases. Ann Biomed Eng. 2011; 39(2):621-35 [link]