Abstract Bioinformatics is a transdisciplinary field based on statistics, machine learning, computer science, biology, and medicine. Computational biology can be considered a subfield of bioinformatics in which computation is analogous to experiments in laboratory science; the emphasis is on the biologist's endpoint (1). There is no standard blueprint for a program in bioinformatics, but perhaps our experience in building the Dept. of Bioinformatics and Computational Biology (DBCB) at UT MD Anderson Cancer Center (MDACC) can offer some instructive ideas, particularly a set of balances that must be struck. Current Personnel (106): • Faculty (15) plus visiting professors • Statistical analysts (17); BioAnalysts (a new job description) (2) • Software engineers in-house (6); on contract (4) • IT staff (4) • Wet laboratory staff (14) • Graduate students (15), postdoctoral fellows (22) • Administrators (7) Essential goals and metrics: • Excellence in academic research: Faculty averaging 9.2 authorships each (in 2016), including 14 overall with IF > 20. • Excellence in Bioinformatics support for projects and programs: In 2016, analysts working on 376 projects for 122 PIs in 50 MDACC departments • Excellence in Education: Teaching in the UT Graduate School of Biomedical Science, etc. • Development of cutting edge bioinformatics software tuned for biologists and clinical researchers (e.g., NGCHMs, MBatch, TCPA, Tanric, MuSE, TransVar, SpliceSeq, VirusSeq, SoS) • Excellence in career development for all: Intensive mentoring Balances to be struck: • Academic research vs. collaborative support: Faculty supervise support, but analysts (17 of them PhDs) do most of it, sharing expertise (and authoring 4.1 authorships each/yr). Faculty's academic projects often flow from collaborative support interactions. • Centralized vs. distributed system for support: We would wish for (a) total immersion of bioinformatics personnel in the basic, translational, or clinical domain being supported; (b) total immersion in the relevant technology and its analytical tools; and (c) total immersion in the bioinformatics community for sharing of expertise, continuing career development, and morale. Clearly, a compromise solution is necessary. Our (imperfect) answer is a hybrid system that emphasizes longitudinal association of faculty and analysts with the biomedical domain department or program supported. • Pure computational vs. combined computational/laboratory research: 5 of 14 faculty have combined programs. There is career benefit but also a cost in time and focus. • Investigator-initiated projects vs. projects that originate from interactions within the institution vs. national/international projects: Our answer has been a self-chosen mix, with deep participation and leadership in NIH projects, notably TCGA. Reference: 1. Weinstein JN. Cancer bioinformatics. In Holland-Frei Cancer Medicine, 9th edition (R. Bast et al., Eds.), 2016. Citation Format: John N. Weinstein, and the Department of Bioinformatics and Computational Biology. Building a program in cancer bioinformatics and computational biology: A balancing act [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 291.