Impact of inherent biases built into proteomic techniques: Proximity labeling and affinity capture compared