Strategies for the multiplex mapping of genes to traits

Abstract

Rewiring and optimization of metabolic networks to enable the production of commercially valuable chemicals is a central goal of metabolic engineering. This prospect is challenged by the complexity of metabolic networks, lack of complete knowledge of gene function(s), and the vast combinatorial genotype space that is available for exploration and optimization. Various approaches have thus been developed to aid in the efficient identification of genes that contribute to a variety of different phenotypes, allowing more rapid design and engineering of traits desired for industrial applications. This review will highlight recent technologies that have enhanced capabilities to map genotype-phenotype relationships on a genome wide scale and emphasize how such approaches enable more efficient design and engineering of complex phenotypes.

Figure 1

Overview of inverse engineering approaches. Inverse engineering begins with a wild-type cell line in which a large number of mutations are introduced in either a random or directed manner to generate phenotypic diversity. These mutants are subjected to screening or selection for a trait of interest (i.e. solvent/temperature tolerance or metabolite production). The pre and post selection populations are sequenced or genotyped by microarray hybridization. The fitness (W_i) of the each genotype (g_i) can be computed by the equation to obtain a plot of fitness (far right circle) where each bar represents a specific genetic loci. Alleles identified by this can be used to guide more focused combinatorial optimization on the trait of interest using directed techniques such as MAGE [16]. While this provides a promising algorithm for directed evolution on the genome scale future technologies will need to be developed to enable this engineering cycle to be preformed rapidly and recursively (indicated by the red arrows) by tracking gene combinations in a high-throughput fashion.

Figure 2

Genome wide fitness measurement techniques. a) Vector library based approaches such as SCALEs [25] start with cloning of random DNA fragments that represent different sites of the genome of interest. This diverse vector library is transformed into the host strain and grown under a defined selective pressure. Following selection the cells are lysed and the vector libraries cut and hybridized to a microarray. The signal intensities from the microarray are then mapped back to the genomic coordinates to identify traits of interest and measure fitness as described in Figure 1. b) Overview of TRMR. An array of rationally designed oligos are cleaved and processed in parallel to produce cassettes with homology arms that guide rational engineering of the entire genome. The TRMR barcodes can then be hybridized to a microarray or PCR amplified and sequenced to determine the fitness of each mutation. This technique provides a versatile method for engineering different functional changes into the genome of interest and rapidly tracking the effects.

Figure 3

Expression profiling to identify adaptive gene expression responses. Changes in gene expression (red line vs. blue) are often correlated with adaptation to changes in the environment or genotypic alterations. Genome wide profiling of such expression changes can thus enable correlations or causal associations to be imputed between the affected genes and the trait of interest. Transcriptome analyses can also define transcriptional units (blue genes) that are co-expressed under varying conditions to allow annotations of unknown gene functions or associations between co-expressed genes.