Underrated Genes

Nature magazine recently published a study of...biological studies. A number of questions were asked, one of them being, “which genes are most represented in the literature?” Not surprisingly, TP53 is the champion, with 9,232 publications. It’s a good read.

A question not addressed is, “What are the most under-represented genes in the literature?” Of course, it’s trivial to find genes that have no mentions at all. What we can do, however, is use our own database and ask, “Which genes have the largest disparity between inclusions in our database and inclusions in the literature?” The exercise is simple on its face, but there are a number of technicalities that make it a bit tricky. If we were writing an academic paper, we’d have to do 100X the work we’re putting into this post. Basically, though, our procedure works like this:  Download a list of genes ranked according to literature mentions. Convert these gene IDs into the format used in the database. Generate a frequency table of all genes in our database. Compare the frequencies in our database against the frequencies in the literature.

The list of genes according to literature mentions is found here: ftp://ftp.ncbi.nih.gov/gene/GeneRIF/generifs_basic.gz

With the understanding that there are a number of ways in which results can be skewed, here’s a list of the most under-rated players in the genomic universe:

RTP4	VSIG2	CLIC6	FAM198B	HIST1H2BD
MT1L	MOXD1	CENPK	ANKRD37	CMBL
PLBD1	TUBA1C	ARHGAP11A	TMEM154	HIST1H2BI
NMES1	TMEM140	PKIA	ADGRL2	KBTBD11
NT5DC2	C15orf15	RSL24D1	RPL27A	FAM49A
PGM5	RGL1	CLMN	EVI2A	TFEC
RPL18A	RPL21	SRM	CALML4	OLFML2A
RPS8	ENDOD1	KDELR3	RPS11	GNG4
TMEM56	SH3BGRL2	CIART	ENPP5	GBP6
RSRP1	COX6A2	GPRIN3	GPRC5C	TMEM71
NRIP3	MFAP3L	CPNE2	ABLIM3	SMIM14
HIST1H2BM	SLC46A3	EVI2B	PCP4L1	TRNP1
GBP4	SLC16A14	RBP7	SLFN13	FAM84A
RAPGEF5	TM6SF1	NSG2	VAT1L	EPPK1
RPL27	DNAJA4	PGAM2	TTC39C	TRANK1
GBP7	N4BP2L2	MEGF6	CDH19	FIBIN
TINAGL1	CCDC3	LONRF2	DDX60L	MXRA7
GPR137B	CENPV	GNG12	CCDC85A	GRAMD3
FAM105A	STRBP	ZNF608	KIAA1551	LRRC2
UAP1L1	MEGF9	EPB41L4A	PLEKHA4	METTL7B

RTP4 is the champion, with few mentions in the literature but more than 700 appearances in our database. Googling RTP4, it seems that there’s no dearth of studies on this gene, but we’re sticking with the above NIH list of literature mentions. Next on the list is VSIG2. A Google search does seem to indicate that nobody cares about this sad gene. It’s hard to even get a clue as to its function.* Nevertheless, it appears 699 times in the database; perturb a cell and there’s a decent chance you’ll alter VSIG2 expression.

We ran a Fisher analysis of an extended, 500-ID list of undervalued genes against our entire database. As might be expected, there’s no massive enrichment for any particular group. There does seem to be a tendency for genes with short transcripts and genes that are depleted in P-bodies to be represented on the list (unadjusted log(P-values) of -7.5 and -5.8). Eyeballing the list, a number of ribosomal proteins can be seen. Perhaps folks view the ribosome as a big unified glob, and don’t care to tinker with its individual components.

The opposite task, that of generating a list of “overrated” genes, is even trickier, and we won’t bother with it here. In the end, genes like TP53 would dominate the list and, given TP53’s role in cancer, labeling it “overrated” or “overstudied” would hardly be fair.

 

*Let’s say you want to know about VSIG2’s function. You can use our tools. First, you enter VSIG2 into our Coregulation app. You’ll get a list of coregulated genes. Take that list and enter it into the Fisher app. To spare you this [minimum] trouble, the swarm of genes with which VSIG2 is coexpressed looks to be hugely involved in the cell cycle, altered by a large array of common drugs (e.g. glucosamine), and also relevant to viral infections. Using the coregulation tool alone, individual genes that are strongly coexpressed with VSIG2 include TRIB3, CHAC1, ASNS, and many more. You can also note that CA9, FAM111B, NREP, and more have a fairly strong tendency to be expressed in the opposite direction to VSIG2 (i.e. when VSIG2 is up, CA9 tends to be down).

whatismygene.com 

Become a Patron!