Navigating gene networks

Statistical analysis

The second part of the summer job project was the statistical analysis of the difference between genes known to be involved in genetic disorders ('disease genes') and all other genes ('non-disease genes', although there could be, and probably are, disease genes in there that are not yet known). This was done by using several databases, mainly two (ENSEMBL and ASTD) for the gene lengths, exon numbers, etc. and two (OMIM and GAD) for the known disease genes.

For all the values, the median of the set was taken. P-values were calculated using the ranksum Matlab command (from the Statistics Toolbox), performing a Mann-Whitney U-test. This tests two data sets for the assumption that they come from an identical distribution. An additional advantage of this test is the possibility to compare data sets of a different sizes.

General comparison

ENSEMBL and OMIM
	Disease genes	Non-disease genes	P-value
Gene length (real)	35384	5856	< 0.001	boxplot
Exon length	133	130	< 0.001	boxplot
CDS length	109	95	< 0.001	boxplot
Gene length (sum of all exon lengths per gene)	5554	8511	< 0.001	boxplot
Gene length (CDS)	3876	5008	< 0.001	boxplot
Number of exons	22	42	< 0.001	boxplot
Number of coding exons	26	39	< 0.001	boxplot

ENSEMBL and GAD
	Disease genes	Non-disease genes	P-value
Gene length (real)	17972	5474	< 0.001	boxplot
Exon length	136	129	< 0.001	boxplot
CDS length	110	99	< 0.001	boxplot
Gene length (sum of all exon lengths per gene)	4542	6515	< 0.001	boxplot
Gene length (CDS)	2180	3239	< 0.001	boxplot
Number of exons	18	31	< 0.001	boxplot
Number of coding exons	15	24	< 0.001	boxplot

ASTD and OMIM
	Disease genes	Non-disease genes	P-value
Exon length	125	127	< 0.001	boxplot
CDS length	116	116	0.4878	boxplot
Intron length	1391	1479	0.0017	boxplot
Number of exons	27	21	< 0.001	boxplot
Number of coding exons	19	13	0.0012	boxplot
Number of introns	22	17	< 0.001	boxplot
Exon length (seq)	125	128	< 0.001	boxplot
CDS length (seq)	129	133	< 0.001	boxplot
Exon CG%	0.5208	0.5208	0.5312	boxplot
CDS CG%	0.5139	0.5124	0.2538	boxplot
Transcription length	747	740	0.1867	boxplot
Transcription CG%	0.5294	0.5313	0.2686	boxplot

ASTD and GAD
	Disease genes	Non-disease genes	P-value
Exon length	134	127	< 0.001	boxplot
CDS length	128	115	< 0.001	boxplot
Intron length	1276	1507	< 0.001	boxplot
Number of exons	20	21	< 0.001	boxplot
Number of coding exons	13	14	< 0.001	boxplot
Number of introns	16	17	< 0.001	boxplot
Exon length (seq)	134	128	< 0.001	boxplot
CDS length (seq)	139	132	< 0.001	boxplot
Exon CG%	0.5379	0.5195	< 0.001	boxplot
CDS CG%	0.5250	0.5109	< 0.001	boxplot
Transcription length	742	739	< 0.001	boxplot
Transcription CG%	0.5437	0.5306	< 0.001	boxplot

Homology

In this part, the resemblance of the human genes in different species is compared between disease and non-disease genes. A distinct difference is visible between disease and non-disease genes. It seems that disease genes are more common than others for all the different organisms studied. This is stastically confirmed by a ranksum test in Matlab, as explained above. All of the homology differences pass the test, i.e. reject the hypothesis that the two samples have the same median. The difference is less pronounced when the test is done using the GAD database of disease genes. Probably, the reason for this is the larger number of 'false positive' disease genes.

Homology - OMIM
	Disease genes	Non-disease genes	P-value
M.musculus - mouse	86.76	53.76	< 0.001	boxplot	histograms
R.norvegicus - rat	82.76	39.76	< 0.001	boxplot
D.melanogaster - fruit fly	0	0	< 0.001	boxplot
G.gallus - chicken	62.76	0	< 0.001	boxplot
P.troglodytes - chimpanzee	98.76	93.76	< 0.001	boxplot	histograms
C.familiaris - dog	85.76	20.76	< 0.001	boxplot
A.gambiae - mosquito	0	0	< 0.001	boxplot
C.elegans - worm	0	0	< 0.001	boxplot	histograms
T.rubripes - fish	45.76	0	< 0.001	boxplot
O.latipes - fish	47.76	0	< 0.001	boxplot
X.tropicalis - frog	52.76	0	< 0.001	boxplot	histograms
D.rerio - zebrafish	42.76	0	< 0.001	boxplot

Homology - GAD
	Disease genes	Non-disease genes	P-value
M.musculus - mouse	79.76	43.76	< 0.001	boxplot	histograms
R.norvegicus - rat	77.26	22.76	< 0.001	boxplot
D.melanogaster - fruit fly	0	0	< 0.001	boxplot
G.gallus - chicken	51.26	0	< 0.001	boxplot
P.troglodytes - chimpanzee	97.76	94.26	< 0.001	boxplot	histograms
C.familiaris - dog	80.26	47.26	< 0.001	boxplot
A.gambiae - mosquito	0	0	< 0.001	boxplot
C.elegans - worm	0	0	< 0.001	boxplot	histograms
T.rubripes - fish	37.76	0	< 0.001	boxplot
O.latipes - fish	38.76	0	< 0.001	boxplot
X.tropicalis - frog	41.76	0	< 0.001	boxplot	histograms
D.rerio - zebrafish	35.76	0	< 0.001	boxplot

Floating base

The floating base characteristics of disease and non-disease genes are compared. No real difference is notable. For the three cases in which there are only two possibilities for the floating gene, the pie chart legend is not correct. The leftmost piece of the pie symbolizes the alphabetically first base. This can also be seen in the table directy underneath.

For all of the fb-comparisons with four possibilities, the p-value is 0.028571. For those with only two possible bases, the p-value is 0.65714.

Legend for the pie charts

Floating base - OMIM
		A	C	G	T
Ala/A - GC*	Disease genes	0.26896	0.31669	0.12889	0.28546	pie chart
	Non-disease genes	0.26944	0.31588	0.12861	0.28608
Gly/G - GG*	Disease genes	0.31287	0.27498	0.25165	0.1605	pie chart
	Non-disease genes	0.31158	0.27353	0.25468	0.16021
Pro/P - CC*	Disease genes	0.30003	0.27487	0.13649	0.28861	pie chart
	Non-disease genes	0.29825	0.27826	0.1321	0.29139
Thr/T - AC*	Disease genes	0.31832	0.30561	0.12054	0.25554	pie chart
	Non-disease genes	0.32206	0.29997	0.11699	0.26097
Val/V - GT*	Disease genes	0.14962	0.23922	0.38326	0.2279	pie chart
	Non-disease genes	0.15288	0.23485	0.38482	0.22745
Arg/R - CG*	Disease genes	0.21763	0.28511	0.3181	0.17915	pie chart
	Non-disease genes	0.20851	0.27878	0.33227	0.18044
Arg/R - AG*	Disease genes	0.52533		0.47467		pie chart
	Non-disease genes	0.53462		0.46538
Leu/L - CT*	Disease genes	0.12396	0.25945	0.39314	0.22345	pie chart
	Non-disease genes	0.12566	0.25288	0.39201	0.22944
Leu/L - TT*	Disease genes	0.39864		0.60136		pie chart
	Non-disease genes	0.39373		0.60627
Ser/S - TC*	Disease genes	0.29392	0.3129	0.10342	0.28977	pie chart
	Non-disease genes	0.29372	0.31249	0.09813	0.29567
Ser/S - AG*	Disease genes		0.60957		0.39043	pie chart
	Non-disease genes		0.59773		0.40227

Floating base - GAD
		A	C	G	T
Ala/A - GC*	Disease genes	0.25955	0.32354	0.12595	0.29096	pie chart
	Non-disease genes	0.27018	0.31544	0.1284	0.28599
Gly/G - GG*	Disease genes	0.3071	0.27557	0.24901	0.16831	pie chart
	Non-disease genes	0.31212	0.27339	0.25422	0.16027
Pro/P - CC*	Disease genes	0.29034	0.28983	0.12814	0.29169	pie chart
	Non-disease genes	0.29855	0.27802	0.13201	0.29142
Thr/T - AC*	Disease genes	0.30966	0.31596	0.11925	0.25512	pie chart
	Non-disease genes	0.3227	0.29946	0.11638	0.26145
Val/V - GT*	Disease genes	0.13994	0.23979	0.40417	0.2161	pie chart
	Non-disease genes	0.15378	0.23432	0.38316	0.22874
Arg/R - CG*	Disease genes	0.2047	0.29708	0.30952	0.1887	pie chart
	Non-disease genes	0.20956	0.27749	0.33277	0.18017
Arg/R - AG*	Disease genes	0.52541		0.47459		pie chart
	Non-disease genes	0.53528		0.46472
Leu/L - CT*	Disease genes	0.12355	0.25177	0.40019	0.22449	pie chart
	Non-disease genes	0.12596	0.25293	0.3909	0.23022
Leu/L - TT*	Disease genes	0.36506		0.63494		pie chart
	Non-disease genes	0.3951		0.6049
Ser/S - TC*	Disease genes	0.29069	0.32631	0.095449	0.28755	pie chart
	Non-disease genes	0.29425	0.31144	0.097969	0.29633
Ser/S - AG*	Disease genes		0.60237		0.39763	pie chart
	Non-disease genes		0.5976		0.4024

laatst aangepast op 14 september 2007