Package 'CoCiteStats'

Compute actor size adjustment on a two way table

Description

When two objects are related through a bipartite graph it is sometimes appropriate to carry out special adjustments. One of the adjustments is called actor size adjustment. In this case the counts are adjusted according to how often the objects are referenced.

Usage

actorAdjTable(twT, eps = 1e-08)

Arguments

twT	A two way table as produced by twowayTable.
eps	A small quantity used to assess approximate equality.

Details

When testing for associations between entities, the social networks literature has developed a number of tools to help measure such associations. We can think of genes (actors) as being joined by citation in papers (events) and having two genes cited in the same paper (equivalent to two actors attending the same event) suggests that they are related to each other. However, some genes are cited in many papers and so we might want to discount the level of importance, as compared to genes that are cited less often. And additionally, some papers cite very many genes, and hence typically say less about them than a paper that cites rather fewer genes.

Value

An adjusted two way table, with elements named u11, u12, u21 and u22.

Author(s)

R. Gentleman

References

Testing Gene Associations Using Co-citation, by B. Ding and R. Gentleman. Bioconductor Technical Report, 2004

See Also

paperLen, twowayTable

Examples

tw1 = twowayTable("10", "100", FALSE)
  actorAdjTable(tw1)

---

Compute gene-gene statistics

Description

Computes gene gene statistics.

Usage

gene.gene.statistic(g1, g2, paperLens = paperLen())

Arguments

g1	The Entrez Gene identifier for one of the genes.
g2	The Entrez Gene identifier for the other gene.
paperLens	A vector with the number of citations for each paper.

Details

For the two genes identified by their Entrez IDs a number of two-way table statistics, i.e. those computed via twTStats, are returned, as are their gene and paper size adjusted variants.

Value

A list with entries

original	The output of twTStats on the observed data.
gs	The output of twTStats on the data scaled for gene size.
ps	The output of twTStats on the data scaled for paper size.
both	The output of twTStats on the data scaled for both paper and gene size.

Author(s)

B. Ding and R. Gentleman

References

Testing Gene Associations Using Co-citation, by B. Ding and R. Gentleman. Bioconductor Technical Report, 2004

See Also

twowayTable

Examples

g1 = "10"  #Entrez ID for gene 1
 g2 = "101"  #Entrez ID for gene 2
 pLens = paperLen()
 gene.gene.statistic(g1, g2, pLens)

---

Compute 3 two-way table statistics and p-values, with 4 different adjustments, for PubMed co-citation

Description

This function calculates Concordance, Jaccard's index and Hubert's $\Gamma$ with no adjustment, adjusting for paper size (PS), adjusting for gene size (GS) and both, to evaluate the significance of co-citation of a gene of interest and a gene list

Usage

gene.geneslist.sig(gene, geneslist, paperLens = paperLen(), n.resamp=100)

Arguments

gene	The Entrez Gene ID for the gene of interest.
geneslist	The list of Entrez Gene IDs for genes with which the co-citation of the gene of interest is to be evaluated.
paperLens	The sizes of the PubMed papers for consideration.
n.resamp	Number of resampling for generating empirical p-values.

Value

Statistics and resampling p-values for all 3 two-way tables along with the 4 adjustments for gene and geneslist based on n.resamp resamplings.

Author(s)

Beiying Ding

References

Testing Gene Associations Using Co-citation, by B. Ding and R. Gentleman. Bioconductor Technical Report, 2004

See Also

actorAdjTable, paperLen, twTStats, twowayTable

Examples

gene <- "705"
geneslist <- "7216"

gene.geneslist.sig(gene, geneslist, n.resamp=50)

---

A function to compute association measures between a gene of interest and a list of genes.

Description

Whether or not a gene has an association with another gene, or a set of genes is measured using co-citation in PubMed as a basis for measuring that association.

Usage

gene.geneslist.statistic(gene, geneslist, paperLens = paperLen())

Arguments

gene	The Entrez Gene ID for the gene of interest.
geneslist	A vector of Entrez Gene ID for the set of genes.
paperLens	A vector containing the number of genes cited by each paper.

Value

To be filled in later.

Author(s)

R. Gentleman

References

Testing Gene Associations Using Co-citation, by B. Ding and R. Gentleman. Bioconductor Technical Report, 2004

See Also

twowayTable, link{gene.gene.statistic}

Examples

g1 = "101"
  gl = c("10014", "10015", "10016", "10017", "10018")
  pL = paperLen()
  s1 = gene.geneslist.statistic(g1, gl, pL)
  s1

---

Find the number of papers cited

Description

The set of papers that cite the input Entrez Gene identifiers are found, and for each of these the number of genes cited in that paper is computed and returned.

Usage

paperLen(x)

Arguments

x	A vector of Entrez Gene identifiers.

Details

This function first finds the set of unique PMIDs associated with the input set of Entrez Gene IDS. Then for each PMID it finds the number of Entrez Gene identifiers associated with that paper. The function uses different sets of variable mappings from the org.Hs.eg.db package.

If x is missing then all Entrez gene identifiers in the org.Hs.egPMID are used.

Value

counts	For each paper the number of Entrez Gene identifiers referred to.
papers	A list of the same length as x, each element contains the papers that refer to the corresponding Entrez Gene identifier.

Author(s)

R. Gentleman

See Also

twTStats

Examples

ans = paperLen(c("10", "1001"))
  ans$counts
  ans$papers

---

Compute a two way co-citation table for 2 genes.

Description

This function computes a two way table for comparing co-citation, in PubMed for the two input genes. The values in the table can be adjusted according to either the paper size or the gene size.

Usage

twowayTable(g1, g2, weights = TRUE, paperLens=paperLen())

Arguments

g1	The EntrezGene identifier for gene 1.
g2	The EntrezGene identifier for gene 2.
weights	TRUE or FALSE indicating whether paper size weights should be used.
paperLens	A vector containing the number of genes each paper refers to, or cites.

Details

To determine the association between two genes one can use co-citation in the medical literature. When weights is FALSE this function computes the number of papers that cite only gene 1, only gene 2, both and neither.

By default, we use the org.Hs.eg.db package to define the set of papers that are used in the computations. For other organisms, or for more restricted sets of papers the user will need to supply the vector paperLens explicitly.

One can consider papers which cite many genes to be less informative than those that cite only a few genes. If weights is TRUE (the default) then papers are weighted by the inverse of the number of citations.

Value

A vector of length four, with entries n11, n12, n21 and n22. These correspond to the number of papers that cite both genes, the number that cite only gene 1, the number that cite only gene 2, and the total number of papers minus those counted in n11, n21, n12, or in the default case the weighted versions of these quantities.

Author(s)

R. Gentleman

See Also

paperLen, twTStats

Examples

pL = paperLen()
  twowayTable("10", "100", paperLens=pL)
  twowayTable("10", "100", FALSE, paperLens=pL)

---

Computes 3 two way table statistics.

Description

For two way tables based on co-citations four different test statistics are reported, the odds ratio, the Concordance, the Jaccard index and Hubert's Gamma.

Usage

twTStats(twT)

Arguments

twT	A two way table, as produced by twowayTable.

Details

The entries in the presumed 2 by 2 table are labeled n11, n12, n21, n22, corresponding to the entries in the first row first column, first row second column etc. The odds ratio is the product of n11 and n22 divided by the product of n12 and n21. The Conordance is simply the n11 entry. The Jaccard index is the n11 entry divided by the sum of n11, n12, and n21. Hubert's Gamma is slightly more complicated.

Value

Concordance	The concordance statistic.
Jaccard	The Jaccard index.
Hubert	Hubert's Gamma
OddsRatio	The odds ratio.

Author(s)

R. Gentleman

References

Testing Gene Associations Using Co-citation, by B. Ding and R. Gentleman. Bioconductor Technical Report, 2004

See Also

paperLen, twowayTable

Examples

tw1 = twowayTable("10", "101", FALSE)
  twTStats(tw1)

---