We present a method to automatically plan a robotic process to mix individual combinations of reactants in individual reaction vessels (vials or wells in a multiwell plate), mixing any number of reactants in any desired stoichiometry, and ordering the mixing steps according to an arbitrarily complex treelike assembly protocol. This process enables the combinatorial generation of complete or partial product libraries in individual reaction vessels from intermediates formed in the presence of different sets of reactants. It can produce either libraries of chimeric genes constructed by ligation of fragments from different parent genes or libraries of chemical compounds constructed by convergent synthesis. Given concentrations of the input reactants and desired amounts or volumes of the products, our algorithm, RoboMix, computes the required reactant volumes and the resulting product concentrations, along with volumes and concentrations for all intermediate combinations. It outputs a sequence of robotic liquid transfer steps that ensures that each combination is correctly mixed even when individualized stoichiometries are employed and with any fractional yield for a product. It can also account for waste in robotic liquid handling and residual volume needed to ensure accurate aspiration. We demonstrate the effectiveness of the method in a test mixing dyes with different UV-vis absorption spectra, verifying the desired combinations spectroscopically.

Oxidation of monoclonal antibodies (mAb) is a common chemical modification with potential impact on a therapeutic protein's activity and immunogenicity. In a previous study, it was found that tryptophan oxidation (Trp-ox) levels of two mAb produced in Chinese hamster ovary (CHO) cells were significantly lowered by modifying cell culture medium/feed. In this study, transcriptome analysis by RNA-Seq is applied to further elucidate the underlying mechanism of those changes in lowering the Trp-ox levels. Cell samples from the 5L fed-batch conditions are harvested and subjected to RNA-Seq analysis. The results showed that the cell culture changes had little impact on neither the expression of the mAb transgenes nor genes related to glycosylation. However, those changes did significantly alter the expression of multiple genes (p-value ≤0.05 and absolute fold change ≥1.5 or adjusted p-value ≤0.1) involved in transport of copper, regulation of glutathione, iron storage, heme reduction, oxidative phosphorylation, and Nrf2-mediated antioxidative response. These findings suggest a key underlying mechanism in lowering Trp-ox levels by CDM was likely to be collectively controlling ROS levels through regulation of those genes' expression. This is the first example, to our knowledge, applying transcriptomic analysis to mechanistically understand the impact of cell culture on mAb oxidation.

The functions of a gene are traditionally annotated textually using either free text (Gene Reference Into Function or GeneRIF) or controlled vocabularies (e.g., Gene Ontology or Disease Ontology). Inspired by the latest word cloud tools developed by the Information Visualization Group at IBM Research, we have prototyped a visual system for capturing gene annotations, which we named Gene Graph Into Function or GeneGIF. Fully developing the GeneGIF system would be a significant effort. To justify the necessity and to specify the design requirements of GeneGIF, we first surveyed the end-user preferences. From 53 responses, we found that a majority (64%, p < 0.05) of the users were either positive or neutral toward using GeneGIF in their daily work (acceptance); in terms of preference, a slight majority (51%, p > 0.05) of the users favored visual presentation of information (GeneGIF) compared to textual (GeneRIF) information. The results of this study indicate that a visual presentation tool, such as GeneGIF, can complement standard textual presentation of gene annotations. Moreover, the survey participants provided many constructive comments that will specify the development of a phase-two project (http://128.248.174.241/) to visually annotate each gene in the human genome. KeywordsGene function; Social networking; Visualization; Word cloud IntroductionGenes in the human genome have been predominantly annotated using unstructured text. For example, the Gene Reference Into Function (GeneRIF) provides a tool to include one or more 255-character-long "gene function" statements that couple a specific publication with a gene [4,6]. An example GeneRIF annotation of the human Kruppel-like factor 4 (KLF4, GeneID:9314) gene is shown in Fig. 79.1. For genes with more than about ten GeneRIFs, it is time-consuming to review the knowledge present in GeneRIFs.Gene Ontology annotations [3] and Disease Ontology annotations [5] of a gene are more compact and the ontological structure makes these annotations much easier for a human reader to parse, in addition to the advantages of these ontologies for semantic reasoning and inference. However, these ontological systems require training to use consistently and Correspondence to: Warren A. Kibbe, wakibbe@northwestern.edu. NIH Public AccessAuthor Manuscript Adv Exp Med Biol. Author manuscript; available in PMC 2011 February 15. NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author Manuscript accurately, and require a significant investment in curatorial time to build the ontological structure.We investigated a different approach to present the genome annotation data. Research in human cognition has suggested that visual presentation can facilitate human learning and knowledge acquisition [2,7,10]. New semantic web tools, such as word clouds, appear to be ideally suited for helping people rapidly parse large amounts of textual data. Thus, we explored the impact of using a word cloud visual presentation of gene annotation information using the latest visualization tools...