Machine learning predicts and provides insights into milk acidification rates of Lactococcus lactis.

Lactococcus lactis strains are important components in industrial starter cultures for cheese manufacturing. They have many strain-dependent properties, which affect the final product. Here, we explored the use of machine learning to create systematic, high-throughput screening methods for these properties. Fast acidification of milk is such a strain-dependent property. To predict the maximum hourly acidification rate (Vmax), we trained Random Forest (RF) models on four different genomic representations: Presence/absence of gene families, counts of Pfam domains, the 8 nucleotide long subsequences of their DNA (8-mers), and the 9 nucleotide long subsequences of their DNA (9-mers). Vmax was measured at different temperatures, volumes, and in the presence or absence of yeast extract. These conditions were added as features in each RF model. The four models were trained on 257 strains, and the correlation between the measured Vmax and the predicted Vmax was evaluated with Pearson Correlation Coefficients (PC) on a separate dataset of 85 strains. The models all had high PC scores: 0.83 (gene presence/absence model), 0.84 (Pfam domain model), 0.76 (8-mer model), and 0.85 (9-mer model). The models all based their predictions on relevant genetic features and showed consensus on systems for lactose metabolism, degradation of casein, and pH stress response. Each model also predicted a set of features not found by the other models.

Targeting Repeated Regions Unique to a Gene Is an Effective Strategy for Discovering Potent and Efficacious Antisense Oligonucleotides.

Well-validated strategies for discovering potent and efficacious antisense oligonucleotides are central to realize the full therapeutic potential of RNA therapy. In this study, we focus on RNA targets where the same sequence of 16-20 nt is found in several regions across the RNA, and not in any other RNA. Targeting such unique repeated regions with oligonucleotides designed as gapmers and capable of recruiting RNase H has previously been proposed as a strategy for identifying potent gapmers. By sequence analysis of the human and monkey transcriptomes, we find that such unique repeated regions in RNA are often conserved between humans and monkeys, which allow pharmacodynamic effects to be evaluated in non-human primates before testing in humans. For eight potential RNA targets chosen in an unbiased fashion, we targeted their unique repeated regions with locked nucleic acid (LNA)-modified gapmers, and for six of them we identified gapmers that were significantly more potent and efficacious in vitro than non-repeat-targeting gapmer controls. We suggest a stochastic model for repeat-targeting gapmers that explains all effects observed so far and can help guide future work. Our results support the targeting of repeated regions as an effective strategy for discovering gapmer antisense oligonucleotides suitable for therapeutic development.