January 2020

SPIED3 is a web tool designed to facilitate fast and simple quantitative querying of publicly available gene expression data. The web tool is an extension of SPIEDw (version 2) and SPIED (version 1), featuring an expanded set of array samples and new compound expression profile databases. SPIED3 replaces the original web tool, which can be accessed here SPIEDw.

Querying SPIED3
SPIED3 is searched with query profile files consisting of a list of genes and usually information about their expression changes. Queries are performed by dropping gene lists into the query box. The first column of the query should contain the gene names and the second column the corresponding expression changes, if these are known. Example query expression profiles: profileX, profileY. If the expression changes are continuous then the scoring statistic is based upon a regression analysis. Otherwise, if only the direction of change is given (+/-1 in the second column) or no expression change values are given then the scoring is based on a Fisher statistic. Gene names are mapped to human equivalents from the HUGO Gene Nomenclature Committee (HGNC, www.genomes.org) database.

SPIED3 datasets and output
The correlation of the query profile with the given SPIED3 entry is ranked according to the Z-score corresponding to a Pearson regression analysis for continuous expression profile queries or the probability score of a cumulative binomial distribution with gene probabilities estimated from database frequencies for discrete expression queries. The output lists significantly correlating SPIED3 entries ranked according to the score. The output format consists of the series id, the sample id, the regression score and the significance. Each output entry also has a web link to the NCBI GEO pages for the given series and samples and a 'magnifying glass' link button to access the query scores against every sample in the series. This functionality enables the user to assess whether there is a correlation between the skew of the enrichment and the given treatment or condition being assayed in the series. For example, if the query corresponds to the expression changes upon treatment with a given drug and this query picks out another instance of the same drug in SPIED3, then one expects there to be a positive enrichment skew with the drug treatment samples and a negative skew with the control samples. The search can be restricted to subsets of SPIED3 based on species.

In addition, SPIED3 hosts the following drug treatment profiles:

DRUGScmap The Connectivity Map 2.0 dataset of 1,309 drug like compounds profiled on human cancer cell lines.

DRUGSlincs The NCBI GEO deposited LINCS L1000 data (GEO accessions GSE70138 and GSE92742) consisting of 15,646 drug like compounds profiled in 98 human cell lines.

DRUGSdm The rat toxicology profiles from the DrugMatrix data (GEO accession GSE59927).

Also, hosted is the LINCS gene perturbation data accessible through GENESlincs.

A further database is based on co-expression hubs, GENEthubs. Here, each gene is mapped to a set of genes that have either strong positive or negative correlations in expression across the totality of the SPIED expression profiles. A strong query correlation can highlight transcription factors underlying the query.

All queries should be sent to:

Gareth Williams
Wolfson Centre for Age-Related Diseases
King's College London
London SE1 1UL

SPIEDw Search

To start SPIEDw search, first enter a query gene expression profile:

The search can be restricted based on species type or restricted to drug treatments only:

If required, the number of output results can be increased beyond the default 100:

Gareth Williams 2023