Abstract
A main challenge of data-driven sciences is how to make maximal use of the progressively expanding databases of experimental datasets in order to keep research cumulative. We introduce the idea of a modeling-based dataset retrieval engine designed for relating a researcher's experimental dataset to earlier work in the field. The search is (i) data-driven to enable new findings, going beyond the state of the art of keyword searches in annotations, (ii) modeling-driven, to include both biological knowledge and insights learned from data, and (iii) scalable, as it is accomplished without building one unified grand model of all data. Assuming each dataset has been modeled beforehand, by the researchers or automatically by database managers, we apply a rapidly computable and optimizable combination model to decompose a new dataset into contributions from earlier relevant models. By using the data-driven decomposition, we identify a network of interrelated datasets from a large annotated human gene expression atlas. While tissue type and disease were major driving forces for determining relevant datasets, the found relationships were richer, and the model-based search was more accurate than the keyword search; moreover, it recovered biologically meaningful relationships that are not straightforwardly visible from annotations—for instance, between cells in different developmental stages such as thymocytes and T-cells. Data-driven links and citations matched to a large extent; the data-driven links even uncovered corrections to the publication data, as two of the most linked datasets were not highly cited and turned out to have wrong publication entries in the database.
Highlights
Molecular biology, historically driven by the pursuit of experimentally characterizing each component of the living cell, has been transformed into a data-driven science [1,2,3,4,5,6] with just as much importance given to the computational and statistical analysis as to experimental design and assay technology
We ask what could be done with these models towards cumulatively building knowledge from data in molecular biology? Speaking about models generally and assuming the many practical issues can be solved technically, we arrive at our answer: we propose creating a modeling-driven dataset retrieval engine, which a researcher can use for positioning her own measurement data into the context of the earlier biology
We benchmarked the combination model against state-of-the-art dataset retrieval by keyword search, in the scenario in which a user queries with a new dataset against a database of earlier released datasets represented by models
Summary
Historically driven by the pursuit of experimentally characterizing each component of the living cell, has been transformed into a data-driven science [1,2,3,4,5,6] with just as much importance given to the computational and statistical analysis as to experimental design and assay technology This has brought to the fore new computational challenges, such as the processing of massive new sequencing data, and new statistical challenges arising from the problem of having relatively few (n) samples characterized for relatively many (p) variables—the ‘‘large p, small n’’ problem. High-throughput technologies often are developed to assay many parallel variables for a single sample in a run, rather than many parallel samples for a single variable, whereas the statistical power to infer properties of biological conditions increases with larger sample sizes. Most labs are restricted to generating datasets with the statistical power to detect only the strongest effects.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
