**4. Statistical analysis of epigenetics biomarkers**

34 Biomarker

aberrant methylation

61% (25/41); pediatric

adult patients : 39%

adult patients : 31%

adult patients : 42,3%

adult patients : 47,2%

adult patients: 50%

p15INK4 - adult patients :

patients: 32% (15/47) ; p16INK4A - adult patients : 37% (15/41) pediatric patients: 8%

(4/47)

(16/41)

(27/86)

(33/78)

(26/55)

(17/34)

Table 3. Methylation alterations in genes involved in the pathogenesis of MDS.

Role in MDS Pathogenesis / Prognosis

prognosis. (Rodrigues et al.,

Higher risk of leukemic transformation. (Brakensiek et al.,

Associated with unfavorable cytogenetic risk group. Poor prognosis. (Xu et al., 2011)

Advanced stages of

(RAEB, RAEB-t). Disease progression. (Lin et al., 2008)

No statistically differences in low and

high risk, it is relatively early event

(Mori et al., 2011)

in MDS.

Occurs in the various subgroups of MDS with higher incidence in RAEB and RAEB-t. (Solomon et al., 2008)

2010)

2005)

MDS

High incidence in RAEB and RAEB-t. Hypermethylation in both genes are involved in evolution from MDS to AML and confers a poor

Genes Function MDS Incidence of

Members of the cyclindependent kinase (CDK) inhibitor family.

important pathway for inhibiting cell growth. Localization: 9p21

Homotypic cell-cell adhesion protein, involved in cellproliferation. Localization: 16q22

Member of statinduced STAT inhibitor (SSI), also known as a suppressor of cytokine signaling (SOCS) family. Localization: 16p13

DAPK1 is a serine/threonine kinase, a positive mediator of gammainterferon induced programmed cell death. Localization:

9q34.1

3p14.2

apoptosis.

Fragile histidine triad gene member of a superfamily HIT of nucleotide binding proteins. Localization:

Induce cell cycle arrest,

Localization: 1p36

They play an

*p15INK4B*

*p16INK4A*

*CDH-1*  (E-cadherin)

*SOCS-1*  (suppressor of cytokine sinaling)

*DAPK1*  (death associated protein kinase)

*FHIT*  (fragile histidine triad)

*RIZ-1*  (retinoblasto ma proteininteracting zinc finger gene)

Biomarkers have become important tools for diagnosis and treatment of a wide range of illnesses, including cancer. Early detection of cancer through biomarkers will allow for the development of new therapeutic procedures in order to increase survival rate of patients diagnosed with cancer. To help the evaluation of new biomarkers for medical practice, we use statistical methods. In this section, we shall discuss statistical techniques for biomarkers evaluation in the myelodysplastic syndrome (MDS). While the application of the methods presented is on biomarkers in MDS, the content of this section can be applied to any medical research.

The main mathematical concept necessary to understand statistical methods is *probability.* Although earlier work on the subject was done by the Italian mathematician Giralamo Cardano (1501-1576), the investigation of probability as a branch of Mathematics sprang about 1654 with two great French mathematicians: Blaise Pascal (1623-1662) and Pierre Fermat (1601–1665). Both Pascal and Fermat were interested in predicting outcomes in the games of chance popular among the French nobility of the mid-seventeenth century. Of course, we shall not do a discourse on probability. But we need to say that the theory of probability underlies the procedures in inferential statistics, which is very useful to medicine and other disciplines in the health field. In our exposition, we will try to avoid mathematical formulas and theorems.
