Molecular Interaction Search Tool (MIST): an integrated resource for mining gene and protein interaction data

Abstract

Model organism and human databases are rich with information about genetic and physical interactions. These data can be used to interpret and guide the analysis of results from new studies and develop new hypotheses. Here, we report the development of the Molecular Interaction Search Tool (MIST; http://fgrtools.hms.harvard.edu/MIST/). The MIST database integrates biological interaction data from yeast, nematode, fly, zebrafish, frog, rat and mouse model systems, as well as human. For individual or short gene lists, the MIST user interface can be used to identify interacting partners based on protein–protein and genetic interaction (GI) data from the species of interest as well as inferred interactions, known as interologs, and to view a corresponding network. The data, interologs and search tools at MIST are also useful for analyzing ‘omics datasets. In addition to describing the integrated database, we also demonstrate how MIST can be used to identify an appropriate cut-off value that balances false positive and negative discovery, and present use-cases for additional types of analysis. Altogether, the MIST database and search tools support visualization and navigation of existing protein and GI data, as well as comparison of new and existing data.

Proximity labeling reveals novel interactomes in live Drosophila tissue

“…Known ring canal proteins are denoted with a solid line outlining the prey gene circle. Dotted lines indicate previously-established interactions mined from the Molecular Interaction Search Tool (MIST) (Hu et al, 2018).…”

“…Pav::APEX also identified Tum, its known binding partner in the centralspindlin complex (Adams et al, 1998;Tao et al, 2016). To see if any of these prey genes were previously shown to interact with one another, we mined data from the Molecular Interaction Search Tool (MIST, an integrated resource for mining gene and protein interaction data; Hu et al, 2018), and imported the established known interactions into the Cytoscape network map ( Figure 4B, dotted lines connecting prey genes).Taken together, these data support the idea that each RC-APEX bait identified unique interactomes that encompass previously-established interactors, new candidate interactors, and known ring canal proteins.…”

“…Known ring canal genes were manually denoted with a solid outline. The Molecular Interaction Search Tool (MIST) (Hu et al, 2018) tool was used to search for previously-established interactions between all high-confidence prey genes, and positive interactions were indicated with a dotted line.…”

Modulation of Yorkie activity by alternative splicing is required for developmental stability

“…Sashimi plots were created with IGV (Integrative Genomics Viewer, https://igv.org/). Protein-protein interactions within the 105 identified genes were tested using Molecular Interaction Search Tool (MIST; http://fgrtools.hms.harvard.edu/MIST/) using the high confidence threshold which retains only the interactions (direct or indirect) supported by several experimental methods 40 . The number of complexes identified was compared to the distribution of the number of complexes found in 50 sets of 105 random genes, which fits a Poisson distribution (l=2.36).…”

Measuring rank robustness in scored protein interaction networks

“…The effect of error, or noise, on networks, such as missing or misplaced edges, is an established research topic in network science [35], and is often studied via network perturbation [36]. In the context of PINs, stochastic models of noise on network edges have been used to re-score interactions [9], determine optimal score thresholds [37], and have been incorporated into community detection [38,39]. However, these models often rely on assumptions about the behaviour of the error: for example, they may assume that interactions are equally likely to be detected regardless of the properties of the proteins involved.…”