Signaling pathways are a cornerstone of systems biology. Several databases store high-quality representations of these pathways that are amenable for automated analyses. Despite painstaking and manual curation, these databases remain incomplete. We present P ATHL INKER, a new computational method to reconstruct the interactions in a signaling pathway of interest. P ATHL INKER efficiently computes multiple short paths from the receptors to transcriptional regulators (TRs) in a pathway within a background protein interaction network. We use P ATHL INKER to accurately reconstruct a comprehensive set of signaling pathways from the NetPath and KEGG databases.
We show that P ATHL INKER has higher precision and recall than several state-of-the-art algorithms, while also ensuring that the resulting network connects receptor proteins to TRs. P ATHL INKER’s reconstruction of the Wnt pathway identified CFTR, an ABC class chloride ion channel transporter, as a novel intermediary that facilitates the signaling of Ryk to Dab2, which are known components of Wnt/β-catenin signaling.
In HEK293 cells, we show that the Ryk–CFTR–Dab2 path is a novel amplifier of β-catenin signaling specifically in response to Wnt 1, 2, 3, and 3a of the 11 Wnts tested. P ATHL INKER captures the structure of signaling pathways as represented in pathway databases better than existing methods.
P ATHL INKER’s success in reconstructing pathways from NetPath and KEGG databases point to its applicability for complementing manual curation of these databases. P ATHL INKER may serve as a promising approach for prioritizing proteins and interactions for experimental study, as illustrated by its discovery of a novel pathway in Wnt/β-catenin signaling. Our supplementary website at provides links to the P ATHL INKER software, input datasets, P ATHL INKER reconstructions of NetPath pathways, and links to interactive visualizations of these reconstructions on GraphSpace. A major focus in systems biology is the identification of the networks of reactions that guide the propagation of cellular signals from receptors to downstream transcriptional regulators (TRs). Over the past two decades, databases have been developed to store the interactions present in signaling pathways, – facilitating their retrieval for computational analyses. While these databases have been iteratively improved over the years, they are still largely built through extensive and time-consuming manual curation.
Further, the proteins and interactions within the same signaling pathway may vary considerably from one database to another.Inspired by these challenges, we sought to develop a computational approach to automatically reconstruct signaling pathways from a background network of molecular interactions (the interactome). We conceptualized the problem as follows : given as input only the receptors and the transcription factors/regulators (TRs) in a specific signaling pathway, can we analyze the interactome to recover the pathway with high accuracy? Several earlier methods have addressed a computationally similar problem of connecting a set of sources or “causes” (akin to receptors) to a set of targets or “effects” (akin to TRs) through a compact sub-network of the interactome. – However, most of these methods are routinely evaluated on data in budding yeast. To tackle the increased complexity of human signaling pathways, we sought to develop an algorithm with two desirable characteristics. First, the method must be able to compute a reconstruction that captures a large subset of the interactions in the curated signaling pathway.
Ideally, it should have a tunable parameter that smoothly determines the size of the solution. Second, to reflect the process of signal transduction, the receptors must be connected to the downstream TRs in the reconstructed pathway.
Overview of the P ATHL INKER algorithm. Given an interactome, we identify a set of receptors and a set of TRs for a particular curated pathway (e.g., Wnt). We apply P ATHL INKER to reconstruct the pathway, ranking proteins and interactions by their first occurrence in the k shortest paths from any receptor to any TR. Using the curated pathway as a ground truth, we evaluate the performance of P ATHL INKER.
We combine the ranked lists for multiple curated pathways to obtain an aggregate evaluation. We develop P ATHL INKER, an algorithm that satisfies both criteria.
P ATHL INKER finds the k highest scoring paths from any receptor to any TR, where k is a user-defined parameter. As the value of k increases, the solution smoothly increases to capture more interactions in the curated pathways. By design, every interaction in the reconstruction lies on some path from a receptor to a TR. Thus, P ATHL INKER satisfies both criteria for a reconstruction algorithm.We apply P ATHL INKER to a comprehensive set of 15 signaling pathways in the NetPath database and 32 pathways in the KEGG database, both of which are manually curated.
Biological Network Model
Compared with several other approaches, – we show that P ATHL INKER is the only method that can reconstruct this pathway with high recall while also ensuring connectivity between receptors and TRs. To further highlight P ATHL INKER’s effectiveness, we examine results for the Wnt pathway in detail. One of the highest scoring paths computed by P ATHL INKER in the Wnt pathway reconstruction suggests that cystic fibrosis transmembrane conductance regulator (CFTR) and its interactions with receptor-like tyrosine kinase (Ryk) and Dab, mitogen-responsive phosphoprotein, homolog 2 (Dab2), both of which are known members of the Wnt pathway, comprise a novel signaling mechanism from Wnts to β-catenin. We experimentally validate this role for CFTR using loss of function short interfering RNA (siRNA)-based silencing.
We first evaluated the ability of P ATHL INKER and other algorithms to reconstruct a diverse collection of 15 signaling pathways in the NetPath database. We then experimentally validated a novel prediction from P ATHL INKER on the Wnt signaling pathway. Pathway reconstructions from the NetPath database Comparison to other algorithmsWe compared P ATHL INKER with six other network-based algorithms , including shortest path (S HORTESTP ATHS, B OWT IEB UILDER ), random walk with restarts (RWR ), network flow (R ESPONSEN ET ), Steiner forest (PCSF ), ANAT, and a greedy seed-based method (Ingenuity Pathway Analyzer (IPA ).
Brief descriptions of these methods and the user-defined parameters we selected appear in. For each pathway reconstruction, we used the interactions in the NetPath pathway as the set of positives and a subsampled set of interactions not present in the NetPath pathway as the set of negatives. For each algorithm, we aggregated the reconstructions of these pathways to measure the precision and recall ( and ). We observed that ANAT, PCSF, R ESPONSEN ET, S HORTESTP ATHS, and B OWT IEB UILDER achieved values of recall. Evaluation of pathway reconstructions aggregated over 15 NetPath pathways. ( a) Precision and recall of the interactions in pathway reconstructions computed by P ATHL INKER and other algorithms.
( b) Precision and recall of P ATHL INKER and RWR without considering interactions adjacent to the pathway (distance=1). ( c) Distances of each interaction from the pathway for P ATHL INKER and RWR at recalls of 0.2, 0.4, and 0.6.
( d) Rank of receptors (top) and TRs (bottom) in the first 1,000 interactions from P ATHL INKER and RWR reconstructions (rank for all interactions in ). ( e) Median values of precision and recall of P ATHL INKER when oversampling and undersampling receptors and TRs. ( f) Precision and recall of P ATHL INKER when recovering proteins compared with interactions. ( g) Precision and recall of P ATHL INKER when reconstructing 15 NetPath pathways compared with 32 KEGG pathways. RWR, random walk with restarts. To determine the source of the false positive interactions in P ATHL INKER compared with RWR, we asked if the false positives were “close” to the pathway as represented in the NetPath database. First, we recomputed precision of all algorithms after ignoring interactions that involved at least one true positive node in the NetPath pathway (“pathway-adjacent negatives”) before subsampling the negatives.
This modification increased the precision for all the algorithms, with P ATHL INKER clearly dominating all the other methods at values of recall between 0.2 and 0.6. To further investigate this trend, we computed each interaction’s distance from any protein in the pathway, where a distance of zero indicated a true positive and a distance of one indicated a pathway-adjacent negative ( and ). At a recall of 0.2, RWR contained a larger proportion of true positives (purple regions) than P ATHL INKER, while the proportion of true positives was similar at recall 0.4 and 0.6. However, the larger proportion of interactions that were at a distance of 1 from the pathway (dark blue regions) across all three values of recall indicates that P ATHL INKER’s false positives were closer to the pathway than RWR’s false positives.To compare P ATHL INKER and RWR using the criterion where we required receptors and TRs to be connected in the reconstruction, we assessed how quickly P ATHL INKER and RWR recovered the curated receptors and TRs. For P ATHL INKER and RWR, we recorded the index of the first interaction that contained each receptor or each TR.
Systems Biology For Signaling Networks Pdf 2017
Shows the results for the first 1,000 ranked interactions, and shows the full ranking. P ATHL INKER and RWR recovered receptors at about the same rate, although P ATHL INKER’s long tail indicated that the last few receptors were difficult for P ATHL INKER to retrieve. Conversely, P ATHL INKER successfully recovered 90% of the TRs in the pathways in the first 1,000 ranked interactions, compared with only 38% recovered by RWR. Evaluation of P ATHL INKER’s performanceWe assessed P ATHL INKER’s performance in several additional ways to investigate its robustness to the inputs and its effectiveness for other pathway databases. First, we added (incorrect) receptors/TRs to the input or removed correct receptors/TRs from the input and compared the resulting reconstructions ( and ).
When we deleted 30% of the receptors and 30% of the TRs from the input, the mean precision at recall of 0.3 and 0.6 dropped by 11% (from 0.42 to 0.38) and 27% (from 0.28 to 0.22), respectively, compared with the precision values with the correct inputs. The results were similar for random additions of 30% of the receptors and 30% of the TRs.Second, we evaluated the performance of recovering proteins in the reconstructions. At similar values of recall, P ATHL INKER’s precision for protein recovery was much higher than that for interaction recovery.
In fact, the precision values of all algorithms improved considerably (comparing with ). When excluding proteins that have an interaction with at least one protein in the pathway, all algorithms have nearly perfect precision.Our analysis thus far relied on 15 pathways from a single database. Our last three assessments estimated the effect of interactions present only in NetPath and extended the scope of the analysis to a larger set of NetPath pathways and to the KEGG database. First, we estimated the reliance of our reconstructions on NetPath-only interactions by applying P ATHL INKER to an interactome that exclude.