Nowadays, to systematically study changes in response to various conditions of an experiment across genome, various high-throughput techniques like microarray [1, 2], RNA-seq [3], proteomics[4] and metabolomics[5]techniques are readily available. These technologies can easily be used to study changes in response to various diseases in an organism.Given such a detailed genome level measurements in disease and healthy conditions, such data can be readily used to find drugs/drug targets. Here, drug/drug targets are desired which change the disease state to healthy state (Fig.1).
On the other hand, altered metabolism has been seen to be associated with many diseases[6].The manual assembly of set of reactions and associated metabolites in a single network format (called Genome Scale Metabolic models (GSSMs)) has allowed to look at metabolism in normal and diseased states in a systematic way [7,8]. Several studies have integrated high throughput data measuring protein and metabolite levels with GSSMs to give context specific GSSMs such as in cancer [9] or Mycobacterium Tuberculosis (Mtb) [10,11] to find drug targets.
To streamline the above process, a pipeline was developed. Here, first the data is produced form healthy and diseased samples using high throughput technologies as mentioned above (Fig.2: Experimental Samples). Then, such data is integrated with the GSSMs (Fig.2: Biological Network) to give context specific metabolic networks. Context specific networks reflect the metabolic state of disease and healthy state. That is, the flux levels of different metabolic reactions, at a genome level, in different disease and healthy states.
Suchcontext specific networks can then be simulated in a computer (Fig.3: Model Simulation). That is, for example,if the effect of deleting a gene in disease state has to be seen, one can delete that gene from the disease network and see how the flux of all the reactions change. This is possible as the GSSMs are connected networks and steady state flux equations are used to describe the fluxes. This allows looking at effect of a single gene at a genome level.After doing such an ‘in-silico’ gene deletion from disease network, one can look at whether the new fluxmatches the healthy state flux (Fig.3: Rewiring disease state to healthy state). If there is sufficient match, then the deleted gene becomes a suitable drug target candidate.
This ‘in-silico’ gene deletion can be done systematically for all the genes in the network (~1500 genes) and effect can be quantified by checking how well the new altered flux of disease state cells matches the healthy state cells. A high value corresponds to good effect (suitable candidate) and vice versa. This gives a gene ranking where topmost gene, on deletion, changes the disease state to healthy state(Fig.3:Model Application).
Such GSMMs can also be used to do an ‘insilico’ drug ranking. As the drug inhibition data is available in various databases like DrugBank, drug addition is simulated by deleting the corresponding gene form the models which the drug inhibits. This gives a drug ranking (Fig.3: Model Application). Such models also account for a drug inhibiting multiple genes. In these cases, to simulate the effect of addition of such drugs to disease state cells, one deletes the multiple genes simultaneously inhibited by such drugs.
Such drug/drug target ranking can be readily used by researchers working in the field of drug discovery to shortlist set of targets/drugs to work on saving loads of time and costs. Such feature is available to researchers across through our DRUID platform.
[1]Brown PO, Botstein D. Exploring the new world of the genome with DNA microarrays. Nat Genet. 1999;21(1 Suppl):33–7.
[2]Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JC, Trent JM, Staudt LM, Hudson Jr J, Boguski MS, et al. The transcriptional program in the response of human fibroblasts to serum. Science. 1999;283(5398):83–7.
[3]Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B. Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods. 2008;5(7):621–8.
[4]Chandramouli, Kondethimmanahalli, and Pei-Yuan Qian. “Proteomics: challenges, techniques and possibilities to overcome biological sample complexity.” Human genomics and proteomics: HGP 2009 (2009).
[5]Dettmer, Katja, Pavel A. Aronov, and Bruce D. Hammock. “Mass spectrometry‐based metabolomics.” Mass spectrometry reviews 26.1 (2007): 51-78.
[6]DeBerardinis, Ralph J., and Craig B. Thompson. “Cellular metabolism and disease: what do metabolic outliers teach us?.” Cell 148.6 (2012): 1132-1144.
[7] Duarte NC, Becker SA, Jamshidi N, Thiele I, Mo ML, et al. (2007) Global reconstruction of the human metabolic network based on genomic and bibliomic data. Proc Natl Acad Sci U S A 104: 1777-1782
[8] Bordbar A, Monk JM, King ZA, Palsson BO (2014) Constraint-based models predict metabolic and associated cellular functions. Nat Rev Genet 15: 107-120.
[9] Yizhak K, Gaude E, Le Devedec S, Waldman YY, Stein GY, et al. (2014) Phenotype-based cell-specific metabolic modeling reveals metabolic liabilities of cancer. Elife 3
[10 ] Bordbar A, Lewis NE, Schellenberger J, Palsson BO, Jamshidi N (2010) Insight into human alveolar macrophage and M. tuberculosis interactions via metabolic reconstructions. Mol Syst Biol 6: 422
[11] Kumar D, Nath L, Kamal MA, Varshney A, Jain A, et al. (2010) Genome-wide Analysis of the Host Intracellular Network that Regulates Survival of Mycobacterium tuberculosis. Cell 140: 731-743
