Our recent publication utilized publicly available ALS-associated human CNS tissue datasets, as well as the transcriptomic and proteomic data obtained from patient-derived iPSC-differentiated motor neurons (diMNs) from Answer ALS to create meta-analyses in PandaOmics, generating target lists with two levels of novelty for detailed target evaluation.

Potential therapeutic targets for ALS were selected based on their ranking calculated by PandaOmics, consistency of the dysregulated expression across different comparisons, druggability, safety assessment, and clinical trial status, yielding a list of 28 potential candidates. The following table (Table 1) displays candidates whose perturbation induced strong to moderate rescue of C9orf72-mediated eye degeneration in Drosophila. The full list of proposed targets was disclosed in the paper.

To validate whether our targets were relevant to the disease, we used a Drosophila model with genomic editing of C9orf72, the most common genetic risk factor of ALS (Zhang et al., 2015). Most people have two GGGGCC hexanucleotide (G₄C₂) repeat in a non-coding region of C9orf72; however, hundreds to thousands of the G₄C₂ repeats can be observed in patients with ALS. This genetic modification has been manipulated and inserted into various in vivo models, such as mice, worms and flies, to mimic ALS-like phenotypes for research purpose.

Panel A is the scale of magnitude of degeneration in fly eyes expressingC9orf72 repeat extension scored from -4 to 2. The control flies (score 0) exhibited eye degeneration caused by expanded hexanucleotide repeat ((G₄C₂)₃₀) in C9orf72. The positive or negative scores corresponds to an increase or decrease in severity of eye degeneration in the flies. Among the 26 targets with fly models, suppression of 18 targets (69%) using RNA interference (RNAi) strongly or moderately rescued C9orf72-mediated eye degeneration (Panel B, Table 1). More importantly, loss of 7 unreported fly orthologs, corresponding to 8 genes (shown in brackets), strongly rescued neurodegeneration (Panel C).

To summarise, ALS is a severe neurodegenerative disease with ill-defined pathogenesis, calling for urgent developments of new therapeutic regimens. Herein, we applied PandaOmics to identify 28 therapeutic candidates for the disease, in which inhibitions of 18 targets were verified to moderately or strongly rescue eye neurodegeneration in c9ALS Drosophila model. This study offers new insights into how AI speeds up the target discovery process from years to months.