Low-resolution analog-to-digital converters (ADCs) offer a compelling approach to reducing power consumption and hardware complexity in large-scale multiple-input multiple-output (LS-MIMO) systems for next-generation wireless networks. However, aggressive quantization significantly increases detection complexity—particularly for high-order modulation schemes—due to elevated noise and interference. In this paper, we propose a novel framework that jointly detects and decodes signals in LS-MIMO systems employing superposition modulation and low-resolution ADCs. Our approach leverages a message-passing (MP) algorithm that decomposes each high-order symbol into multiple BPSK layers, enabling more tractable soft-interference cancellation and extrinsic information updates. To theoretically assess system performance, we further develop a modified Protograph Extrinsic Information Transfer (PEXIT) algorithm, which integrates the effects of MIMO fading, ADC quantization noise, and protograph LDPC decoding into a single analytical tool. This modified PEXIT accurately predicts the iterative decoding thresholds under varying antenna dimensions, ADC resolutions, and superposition weighting schemes, thus facilitating a systematic code and modulation design process. Extensive numerical results corroborate the analysis, demonstrating that equal-weight superposition modulation provides consistent performance gains over conventional equal-distance constellations, especially in low-to-moderate quantization scenarios. These findings highlight the importance of power allocation strategies for superposition modulation in practical LS-MIMO receivers with coarse quantizers. Overall, this work offers a unified perspective on superposition modulation, low-resolution ADCs, and advanced LDPC decoding, opening new avenues for designing energy-efficient and spectrally efficient communication systems suitable for 5G-and-beyond networks.