[en] In the present thesis several new contributions are made to achieve this goal. Different gauge-invariant subsets of graphs, i.e. whole color factors, are calculated, in order to break the ground for the systematic evaluation of new topologies and to develop corresponding computer algebra codes and computational algorithms to render a part of these problems. Furthermore, also some new results are obtained on the 2-loop level. The work focuses on the heavy quark corrections in the asymptotic region Q² >> m², where the heavy flavor Wilson coefficients factorize into the light flavor Wilson coefficients and massive operator matrix elements. New contributions are obtained for the complete O(α_s³n_fT_F²) corrections to the operator matrix elements A_gq,Q and A_gg,Q. The computation of the Feynman integrals is performed using representations in generalized hypergeometric functions and finite sums. These sums are performed using modern symbolic summation methods implemented in the packages Sigma, Evaluate Multi Sums, and Sum Production. The results are renormalized and checked against Mellin moments. Furthermore, also the 2-loop corrections to the polarized massive OMEs ΔA_gq,Q and ΔA_gg,Q are calculated. Since the calculations are performed in dimensional regularization and Levi-Civita tensors are present in the diagrams, the OMEs are subject to a finite renormalization. New methods are developed to calculate genuine 3-loop topologies of ladder- and V-type, taking into account the number of heavy quark lines involved. The calculation methods involves mapping the Feynman parameterized representations onto multi-sums and using properties of Appell functions and other generalizations of hypergeometric functions. Two integrals are presented, for which the solution with summation methods remains yet an open problem. At three loops, for the first time also graphs with two distinct massive lines occur. A new method is presented for the calculation of such diagrams with equal masses, contributing to the OMEs A_gq,Q and A_gg,Q. The method uses a Mellin-Barnes representation instead of a generalized hypergeometric function and keeps, for convergence reasons, one of the Feynman parameter integrals unintegrated. The above symbolic summation methods are used to solve the sum of residues in terms of cyclotomic harmonic polylogarithms. Many properties of these functions are implemented in the package Harmonic Sums. Since the result is first derived as a generating function, the symbolic summation machinery is applied a second time, solving difference equations and simplifying sums needed to derive the Nth Taylor coefficient for symbolic N. First the O(α_s) contributions are revisited, due to partly different results in the foregoing literature, which can be clarified. At 1-loop order, an efficient representation in Mellin space allowing for fast numerical evaluations is designed, including power corrections. Also here errors in the literature are corrected. Here the 1-loop expressions are also expanded for 1>>m²/Q² up to the constant term. A careful recalculation of the gluonic contribution is performed as well as a calculation in leading logarithmic approximation. The leading logarithmic calculation shows that the same sign error occurs for the pure-singlet contribution at two loops. The heavy quark corrections of charged current deep-inelastic scattering are extended to 2-loop order. The factorization of the heavy flavor Wilson coefficients at large values of Q² is derived for the charged current case. Using the light flavor Wilson coefficients and operator matrix elements up to 2-loop order from the literature, x- and N-space expressions for all heavy flavor Wilson coefficients at two loops are given.