Skin sensitization is one of the key adverse effects to be addressed in the human hazard characterization and risk assessment of chemicals and cosmetic ingredients. Traditionally, the skin sensitization potential of substances has been addressed by using animal models, but an increasing political and ethical demand to replace animal experiments resulted in:
the regulatory requirement in Europe to assess the safety of cosmetic ingredients without testing on animals 
a recent amendment of the industrial chemical regulation explicitly requests information on skin sensitization from non-animal test methods and prescribes that animal testing can only be conducted if the non-animal methods are not applicable or the obtained results are inadequate for classification and risk assessment purposes. 
The OECD has summarized the detailed understanding of the biological mechanism of skin sensitization by structuring the various biological steps in the process into an Adverse Outcome Pathway (AOP) [3,4].
The molecular initiating event (MIE / KE1) is covalent binding to skin proteins (specifically, to cysteine and/or lysine residues) which leads to keratinocytes' activation, a key event (KE2) at cellular level. Another key event at cellular level is activation of dendritic cells (KE3), which is caused by hapten-protein complexes as well as by signalling from activated keratinocytes. Dendritic cells subsequently mature and migrate out of the epidermis to the local lymph node where they display major histocompatibility complex molecules, which include part of the hapten-protein complex to naive T-lymphocytes (T-cells). This induces differentiation and proliferation of allergen chemical-specific memory T-cells (KE4). This signifies the consecutive KE resulting in the acquisition of sensitisation, the adverse outcome on organ level. A sensitised subject has the capacity then to mount a more accelerated secondary response to the same chemical. Thus, if exposure occurs again, at the same or a different skin site, an aggressive immune response will be elicited resulting in allergic contact dermatitis.
Figure adapted from [3,4].
A number of non-animal test methods, in silico, in chemico and in vitro addressing the MIE or one or more of the KEs, are currently at different stages of development.
In silico / computational methods are practical and often inexpensive tools that contribute to predictions of skin sensitiser hazard and potency, identification of read across candidates, prediction of physicochemical properties (e.g. molecular hydrophobicity logP, melting point, boiling point, molecular weight), and identification of whether a chemical is reactive (and if so by which mechanism). Though in silico approaches are not yet sufficiently reliable to assess the skin sensitisation potential or potency of chemicals as stand-alone approaches, confidence in predictions can be strengthened when used along with expert judgement or by incorporation into Defined Approaches (DAs) and/or in a Weight of Evidence (WoE).
add tools / add in silico protocols
In chemico and in vitro methods which have achieved OECD acceptance are: In chemico Skin sensitization: Direct Peptide Reactivity Assay (DPRA), that addresses the MIE by measuring depletion of synthetic peptides ; in vitro Skin sensitization assays addressing the AOP KE1 on keratinocytes activation, that is based on the activation of the Keap1-Nrf2-ARE pathway  and in vitro Skin sensitization assays addressing the AOP KE2 on activation of dendritic cells . For more details you can watch this video . It is a 40 minutes video. In the first 16 minutes Dr Kolle introduces data requirements, including the 2016 update to REACH Annex VII, and provides an overview of the key mechanism of skin sensitisation, based on the published adverse outcome pathway. After that Dr Casati describes the in vitro and in chemico methods that can be used to assess skin sensitisation, with a specific focus on the validated and OECD-adopted methods and on combining them in defined approaches. Dr Kolle provides examples of defined approaches to hazard identification and potency categorisation. Dr Casati closes by reviewing current OECD activities in the field of skin sensitisation.
Cosmetic Europe has set-up a four-step scientific program to achieve regulatory accepted animal free testing strategies that enable the cosmetic industry to conduct skin sensitization safety assessments, if possible by 2020. In the first phase many of the existing test methods were reviewed and evaluated. In the second phase  human data were explored as a reference for assessing the predictivity of individual test methods and testing strategies potentially reducing the uncertainty introduced by the use of animal data in human risk assessment. A database compiled of existing and newly generated data of the test methods together with LLNA and human reference data for 128 substances is presented. The database is described in detail and is used to evaluate the individual test methods’predictivity and aspects of applicability in a harmonized way. These evaluations were complemented  by assessing the twelve DAs and IATA for skin sensitization that have been proposed to the OECD as case studies . The initial database is being expanded to 169 chemicals in Phase III. Ongoing Phase IV activities aim to develop next generation risk assessment (NGRA) case studies and iteratively share those in a series of risk assessment workshops. The first of these workshops was held in summer 2017, followed by two more workshops in 2019. By sharing risk assessments in such a way, it was possible to identify overarching principles describing how to structure a NGRA for skin sensitisation, which are addressed in this publication .
add something on iata and da