Recent progress in analytical calculation of the multiple {inverse, binomial, harmonic} sums , related with ε-expansion of the hypergeometric function of one variable are discussed.1. In the framework of the dimensional regularization [1] many Feynman diagrams can be written as hypergeometric series of several variables [2] (some of them can be equal to the rational numbers). This result can be deduced via MellinBarnes technique [3] or as solution of the differential equation for Feynman amplitude [4]. However, for application to the calculation of radiative corrections mainly the construction of the ε-expansion is interesting. At the present moment, several algorithms for the ε-expansion of different hypergeometric functions have been elaborated. They were mainly related to calculations of concrete Feynman diagrams [5]. Only recently, the general algorithm for integer values of parameters has been described in [6] and its generalization has been done in [7]. The results of expansion are expressible in terms of the new functions, like harmonic polylogarithms [8] or their recent generalization [6,9]. Let us shortly describe, how this algorithm woks on the example of the generalized hypergeometric function of one variable. The starting point is the series representation: P F Q {A 1 +a 1 ε},{A 2 +a 2 ε}, · · · {A P +a P ε} {B 1 +b 1 ε},{B 2 +b 2 ε}, · · · {B Q +b q ε} zwhere (α) j ≡ Γ(α + j)/Γ(α) is the Pochhammer symbol. We concentrate on the case Q+1 F Q , when series converges for all |z| < 1, and on the integer or half-integer values of the parameters {A i , B j } ∈ {m i , m j + 1 2 }. To perform the ε- * Work was supported by DFG under Contract SFB/TR 9-03 and in part by RFBR grant # 04-02-17192.expansion we use the well-known representationwhere m is an integer positive number, m > 1 and S k (j) = j l=1 l −k is the harmonic sum satisfying the relation S k (j) = S k (j − 1) + 1/j k . To work only with positive values for parameters of hypergeometric function we can apply several times the Kummer relation: