A deep multi-task learning approach to identifying mummy berry infection sites, the disease stage, and severity.

Introduction: Mummy berry is a serious disease that may result in up to 70 percent of yield loss for lowbush blueberries. Practical mummy berry disease detection, stage classification and severity estimation remain great challenges for computer vision-based approaches because images taken in lowbush blueberry fields are usually a mixture of different plant parts (leaves, bud, flowers and fruits) with a very complex background. Specifically, typical problems hindering this effort included data scarcity due to high manual labelling cost, tiny and low contrast disease features interfered and occluded by healthy plant parts, and over-complicated deep neural networks which made deployment of a predictive system difficult. Methods: Using real and raw blueberry field images, this research proposed a deep multi-task learning (MTL) approach to simultaneously accomplish three disease detection tasks: identification of infection sites, classification of disease stage, and severity estimation. By further incorporating novel superimposed attention mechanism modules and grouped convolutions to the deep neural network, enabled disease feature extraction from both channel and spatial perspectives, achieving better detection performance in open and complex environments, while having lower computational cost and faster convergence rate. Results: Experimental results demonstrated that our approach achieved higher detection efficiency compared with the state-of-the-art deep learning models in terms of detection accuracy, while having three main advantages: 1) field images mixed with various types of lowbush blueberry plant organs under a complex background can be used for disease detection; 2) parameter sharing among different tasks greatly reduced the size of training samples and saved 60% training time than when the three tasks (data preparation, model development and exploration) were trained separately; and 3) only one-sixth of the network parameter size (23.98M vs. 138.36M) and one-fifteenth of the computational cost (1.13G vs. 15.48G FLOPs) were used when compared with the most popular Convolutional Neural Network VGG16. Discussion: These features make our solution very promising for future mobile deployment such as a drone carried task unit for real-time field surveillance. As an automatic approach to fast disease diagnosis, it can be a useful technical tool to provide growers real time disease information that can prevent further disease transmission and more severe effects on yield due to fruit mummification.

Population Structure and Reproductive Biology of Monilinia vaccinii-corymbosi in Vaccinium angustifolium in Maine.

The fungus Monilinia vaccinii-corymbosi (Mvc) causes mummy berry disease in blueberries including lowbush blueberry, Vaccinium angustifolium, and is a significant pathogen of concern for Maine lowbush blueberry growers. This disease is typically managed with fungicides or by burning of plant debris containing overwintering pseudosclerotia. The population structure of Mvc in various fields in Maine was investigated using microsatellites and isolates collected from three stages in the Mvc lifecycle. The impacts of management strategies were also examined. A high level of genetic diversity was observed in Mvc from 12 lowbush blueberry fields with 199 unique multilocus haplotypes (MLHs) occurring in an original sample of 232 isolates. Twelve private alleles, including six private alleles with frequencies above 0.05, which indicated gene flow, were observed in six out of 12 fields. The population of Mvc in Maine as a whole is mostly a sexual, outcrossing population, as was seen in the diversity of MLHs and low amounts of linkage disequilibrium, although some apothecia appear to result from selfing. Three fields appear to have some clonal reproduction but were not strictly clonal, as multiple MLHs were noted in these fields. Management does not appear to affect population structure, and Mvc may be one large statewide population in Maine.

Evaluating the potential of lignosulfonates and chitosans as alfalfa hay preservatives using in vitro techniques.

Our objectives were to compare the antifungal activity of 5 lignosulfonates, and 2 chitosans against fungi isolated from spoiled hay, and assess the effects of an optimized lignosulfonate, chitosan, and propionic acid (PRP) on high-moisture alfalfa hay. In experiment 1, we determined the minimum inhibitory concentration and minimum fungicidal concentration of 4 sodium lignosulfonates, 1 magnesium lignosulfonate, 2 chitosans, and PRP (positive control) against Aspergillus amoenus, Mucor circinelloides, Penicillium solitum, and Debaromyces hansenii at pH 4 and 6. Among sodium lignosulfonates, the one from Sappi Ltd. (NaSP) was the most antifungal at pH 4. However, chitosans had the strongest fungicidal activity with the exception of M. circinelloides at both pH 4 and 6. PRP had more antifungal effects than NaSP and was only better than chitosans for M. circinelloides. In experiment 2, we evaluated the effects of 3 additives (ADV): optimized NaSP (NaSP-O, UMaine), naïve chitosan (ChNv, Sigma-Aldrich), and PRP on high-moisture alfalfa hay. The experimental design was a randomized complete block design replicated 5 times. Treatment design was the factorial combination of 3 ADV× 5 doses (0, 0.25, 0.5, 1, and 2% w/w fresh basis). Additives were added to 35 g of sterile alfalfa hay (71.5 ± 0.23% DM), inoculated with a mixture of previously isolated spoilage fungi (5.8 log cfu/fresh g), and aerobically incubated in vitro for 23 d (25°C). After incubation, DM losses were reduced by doses as low as 0.25% for both NaSP-O and PRP (x¯=1.61) vs. untreated hay (24.0%), partially due to the decrease of mold and yeast counts as their doses increased. Also, hay NH3-N was lower in NaSP-O and PRP, with doses as low as 0.25%, relative to untreated hay (x¯=1.13 vs. 7.80% of N, respectively). Both NaSP-O and PRP increased digestible DM recovery (x¯=69.7) and total volatile fatty acids (x¯=94.3), with doses as low as 0.25%, compared with untreated hay (52.7% and 83.8 mM, respectively). However, ChNv did not decrease mold nor yeast counts (x¯=6.59 and x¯=6.16 log cfu/fresh g, respectively) and did not prevent DM losses relative to untreated hay. Overall, when using an alfalfa hay substrate in vitro, NaSP-O was able to prevent fungal spoilage to a similar extent to PRP. Thus, further studies are warranted to develop NaSP-O as a hay preservative under field conditions.

In our first experiment, we assessed the antifungal activity of two major types of byproducts, one known as lignosulfonates (5 types), which are generated by paper mills, and another known as chitosans (2 types), which are generated from shellfish. These were tested against four fungi isolated from spoiled hay. We observed that acidic conditions are not necessary for chitosans but are crucial to activate the antifungal properties of lignosulfonates. Also, we found that sodium lignosulfonate from Sappi Ltd. was the most antifungal relative to other sodium lignosulfonates from other manufacturers. Chitosans had stronger fungicidal activity than propionic acid or lignosulfonates against all but one mold tested. In our second experiment, we compared the best treatments from experiment 1 against propionic acid using alfalfa hay as a substrate to grow the same fungi tested in experiment 1. None of the doses of chitosan prevented spoilage on high moisture hay, showing results similar to untreated hay. In contrast, an optimized sodium lignosulfonate and propionic acid prevented fungal spoilage of alfalfa hay with doses as low as 0.25%.

Using peer discussion facilitated by clicker questions in an informal education setting: enhancing farmer learning of science.

Blueberry growers in Maine attend annual Cooperative Extension presentations given by university faculty members. These presentations cover topics, such as, how to prevent plant disease and monitor for insect pests. In 2012, in order to make the sessions more interactive and promote learning, clicker questions and peer discussion were incorporated into the presentations. Similar to what has been shown at the undergraduate level, after peer discussion, more blueberry growers gave correct answers to multiple-choice questions than when answering independently. Furthermore, because blueberry growers are characterized by diverse levels of education, experience in the field etc., we were able to determine whether demographic factors were associated with changes in performance after peer discussion. Taken together, our results suggest that clicker questions and peer discussion work equally well with adults from a variety of demographic backgrounds without disadvantaging a subset of the population and provide an important learning opportunity to the least formally educated members. Our results also indicate that clicker questions with peer discussion were viewed as a positive addition to university-related informal science education sessions.

A DNA-based assay identifies Batrachochytrium dendrobatidis in amphibians.

Chytridiomycosis caused by Batrachochytrium dendrobatidis (Chytridiomycota) has been implicated in declines of amphibian populations on four continents. We have developed a sensitive and specific polymerase chain reaction-based assay to detect this pathogen. We isolated B. dendrobatidis from captive and wild amphibians collected across North America and sequenced the internal transcribed spacer regions of the rDNA cassette of multiple isolates. We identified two primers (Bd1a and Bd2a) that are specific to B. dendrobatidis under amplification conditions described in this study. DNA amplification with Bd1a/Bd2a primers produced a fragment of approximately 300 bp from B. dendrobatidis DNA but not from DNA of other species of chytrids or common soil fungi. The assay detected 10 zoospores or 10 pg of DNA from B. dendrobatidis and detected infections in skin samples from a tiger salamander (Ambystoma tigrinum), boreal toads (Bufo boreas), Wyoming toads (Bufo baxteri), and smooth-sided toads (Bufo guttatus). This assay required only small samples of skin and can be used to process a large number of samples.

Physiology of Batrachochytrium dendrobatidis, a chytrid pathogen of amphibians.

Batrachochytrium dendrobatidis is a pathogen of amphibians that has been implicated in severe population declines on several continents. We investigated the zoospore activity, physiology and protease production of B. dendrobatidis to help understand the epidemiology of this pathogen. More than 95% of zoospores stopped moving within 24 h and swam less than 2 cm before encysting. Isolates of B. dendrobatidis grew and reproduced at temperatures of 4-25 C and at pH 4-8. Growth was maximal at 17-25 C and at pH 6-7. Exposure of cultures to 30 C for 8 d killed 50% of the replicates. B. dendrobatidis cultures grew on autoclaved snakeskin and 1% keratin agar, but they grew best in tryptone or peptonized milk and did not require additional sugars when grown in tryptone. B. dendrobatidis produced extracellular proteases that degraded casein and gelatin but had no measurable activity against keratin azure. The proteases were active against azocasein at temperatures of 6-37 C and in a pH range of 6-8, with the highest activity at temperatures of 23-30 C and at pH 8. The implications of these observations on disease transmission and development are discussed.