No abstract

Vegetative incompatibility in fungi has long been known to reduce the transmission of viruses between individuals, but the barrier to transmission is incomplete. In replicated laboratory assays, we showed conclusively that the transmission of viruses between individuals of the chestnut blight fungus Cryphonectria parasitica is controlled primarily by vegetative incompatibility (vic) genes. By replicating vic genotypes in independent fungal isolates, we quantified the effect of heteroallelism at each of six vic loci on virus transmission. Transmission occurs with 100% frequency when donor and recipient isolates have the same vic genotypes, but heteroallelism at one or more vic loci generally reduces virus transmission. Transmission was variable among single heteroallelic loci. At the extremes, heteroallelism at vic4 had no effect on virus transmission, but transmission occurred in only 21% of pairings that were heteroallelic at vic2. Intermediate frequencies of transmission were observed when vic3 and vic6 were heteroallelic (76 and 32%, respectively). When vic1, vic2, and vic7 were heteroallelic, the frequency of transmission depended on which alleles were present in the donor and the recipient. The effect of heteroallelism at two vic loci was mostly additive, although small but statistically significant interactions (epistasis) were observed in four pairs of vic loci. A logistic regression model was developed to predict the probability of virus transmission between vic genotypes. Heteroallelism at vic loci, asymmetry, and epistasis were the dominant factors controlling transmission, but host genetic background also was statistically significant, indicating that vic genes alone cannot explain all the variation in virus transmission. Predictions from the logistic regression model were highly correlated to independent transmission tests with field isolates. Our model can be used to estimate horizontal transmission rates as a function of host genetics in natural populations of C. parasitica.

Sequence-to-sequence (S2S) pre-training using large monolingual data is known to improve performance for various S2S NLP tasks. However, large monolingual corpora might not always be available for the languages of interest (LOI). Thus, we propose to exploit monolingual corpora of other languages to complement the scarcity of monolingual corpora for the LOI. We utilize script mapping (Chinese to Japanese) to increase the similarity (number of cognates) between the monolingual corpora of helping languages and LOI. An empirical case study of low-resource Japanese-English neural machine translation (NMT) reveals that leveraging large Chinese and French monolingual corpora can help overcome the shortage of Japanese and English monolingual corpora, respectively, for S2S pre-training. Using only Chinese and French monolingual corpora, we were able to improve Japanese-English translation quality by up to 8.5 BLEU in lowresource scenarios.

No abstract

No abstract