Application of Multiple Unsupervised Models to Validate Clusters Robustness in Characterizing Smallholder Dairy Farmers.

The heterogeneity of smallholder dairy production systems complicates service provision, information sharing, and dissemination of new technologies, especially those needed to maximize productivity and profitability. In order to obtain homogenous groups within which interventions can be made, it is necessary to define clusters of farmers who undertake similar management activities. This paper explores robustness of production cluster definition using various unsupervised learning algorithms to assess the best approach to define clusters. Data were collected from 8179 smallholder dairy farms in Ethiopia and Tanzania. From a total of 500 variables, selection of the 35 variables used in defining production clusters and household membership to these clusters was determined by Principal Component Analysis and domain expert knowledge. Three clustering algorithms, K-means, fuzzy, and Self-Organizing Maps (SOM), were compared in terms of their grouping consistency and prediction accuracy. The model with the least household reallocation between clusters for training and testing data was deemed the most robust. Prediction accuracy was obtained by fitting a model with fixed effects model including production clusters on milk yield, sales, and choice of breeding method. Results indicated that, for the Ethiopian dataset, clusters derived from the fuzzy algorithm had the highest predictive power (77% for milk yield and 48% for milk sales), while for the Tanzania data, clusters derived from Self-Organizing Maps were the best performing. The average cluster membership reallocation was 15%, 12%, and 34% for K-means, SOM, and fuzzy, respectively, for households in Ethiopia. Based on the divergent performance of the various algorithms evaluated, it is evident that, despite similar information being available for the study populations, the uniqueness of the data from each country provided an over-riding influence on cluster robustness and prediction accuracy. The results obtained in this study demonstrate the difficulty of generalizing model application and use across countries and production systems, despite seemingly similar information being collected.

Species-informative SNP markers for characterising freshwater prawns of genus Macrobrachium in Cameroon.

Single Nucleotide Polymorphisms (SNPs) are now popular for a myriad of applications in animal and plant species including, ancestry assignment, conservation genetics, breeding, and traceability of animal products. The objective of this study was to develop a customized cost-effective SNP panel for genetic characterisation of Macrobrachium species in Cameroon. The SNPs identified in a previous characterization study were screened as viable candidates for the reduced panel. Starting from a full set of 1,814 SNPs, a total of 72 core SNPs were chosen using conventional approaches: allele frequency differentials, minor allele frequency profiles, and Wright's Fst statistics. The discriminatory power of reduced set of informative SNPs were then tested using the admixture analysis, principal component analysis, and discriminant analysis of principal components. The panel of prioritised SNP markers (i.e., N = 72 SNPs) distinguished Macrobrachium species with 100% accuracy. However, large sample size is needed to identify more informative SNPs for discriminating genetically closely related species, including M. macrobrachion versus M. vollenhovenii and M. sollaudii versus M. dux. Overall, the findings in this study show that we can accurately characterise Macrobrachium using a small set of core SNPs which could be useful for this economically important species in Cameroon. Given the results obtained in this study, a larger independent validation sample set will be needed to confirm the discriminative capacity of this SNP panel for wider commercial and research applications.

Genetic Connectivity of the Sky Emperor, Lethrinus mahsena Populations Across a Gradient of Exploitation Rates in Coastal Kenya.

Marine-protected areas (MPAs) have the potential to enhance fisheries through transport of larvae or by a net emigration of adult and juvenile fish to adjacent fished areas. A network of appropriately located MPAs will have the potential to reseed fished areas and other MPAs. Connectivity studies are therefore important to assess the effectiveness of a network of MPAs and to determine the spatial scale necessary for spillover effects. The principal aim of this study was to determine the potential for Kenyan MPAs to reseed adjacent fishing zones by evaluating the levels of genetic differentiation of populations of Lethrinus mahsena, a commercially important fish, along a continuum of protected and nonprotected sites. Fish samples were collected from MPAs (Mombasa and Kisite Mpunguti Marine Parks) and the fished reserves adjacent to the two MPAs. Total length and weight of the fish from the sites and fin clips from one of the pectoral fins were collected and preserved in 90% ethanol. Genomic profiles for each sample were obtained through genotyping by sequencing using diversity array technology markers. Results from population structure, diversity, and admixture analyses indicated very low genetic differentiation (F ST = 0.00184, P > 0.05) and low population substructure between samples obtained from the study locations implying a free exchange of fish across protected and nonprotected sites. There was a high gene flow and multidirectional migration rate among the sampling sites. Inbreeding was moderately high (F IS = 0.15, P < 0.05) in the marine parks, indicating high relatedness and probably limited mating options for the species due to small population size or spatial restriction. The lack of genetic differentiation between protected areas and open fishing grounds is indicative of genetic connectivity for the sky emperor. This reinforces the significance of maintaining protected areas to serve as breeding and spawning grounds of fish without adversely affecting the livelihoods of communities that depend on the various fisheries linked to MPAs.

Performance Evaluation of Highly Admixed Tanzanian Smallholder Dairy Cattle Using SNP Derived Kinship Matrix.

The main purpose of this study was to understand the type of dairy cattle that can be optimally used by smallholder farmers in various production environments such that they will maximize their yields without increasing the level of inputs. Anecdotal evidence and previous research suggests that the optimal level of taurine inheritance in crossbred animals lies between 50 and 75% when considering total productivity in tropical management clusters. We set out to assess the relationship between breed composition and productivity for various smallholder production systems in Tanzania. We surveyed 654 smallholder dairy households over a 1-year period and grouped them into production clusters. Based on supplementary feeding, milk productivity and sale as well as household wealth status four clusters were described: low-feed-low-output subsistence, medium-feed-low-output subsistence, maize germ intensive semi-commercial and feed intensive commercial management clusters. About 839 crossbred cows were genotyped at approximately 150,000 single nucleotide polymorphism (SNP) loci and their breed composition determined. Percentage dairyness (proportion of genes from international dairy breeds) was estimated through admixture analysis with Holstein, Friesian, Norwegian Red, Jersey, Guernsey, N'Dama, Gir, and Zebu as references. Four breed types were defined as RED-GUE (Norwegian Red/Friesian-Guernsey; Norwegian Red/Friesian-Jersey), RED-HOL (Norwegian Red/Friesian-Holstein), RED-Zebu (Norwegian Red/Friesian-Zebu), Zebu-RED (Zebu-Norwegian Red/Friesian) based on the combination of breeds that make up the top 76% breed composition. A fixed regression model using a genomic kinship matrix was used to analyze milk yield records. The fitted model accounted for year-month-test-date, parity, age, breed type and the production clusters as fixed effects in the model in addition to random effects of animal and permanent environment effect. Results suggested that RED-Zebu breed type with dairyness between 75 and 85% is the most appropriate for a majority of smallholder management clusters. Additionally, for farmers in the feed intensive management group, animals with a Holstein genetic background with at least 75% dairy composition were the best performing. These results indicate that matching breed type to production management group is central to maximizing productivity in smallholder systems. The findings from this study can serve as a basis to inform the development of the dairy sector in Tanzania and beyond.

Morphological and molecular characterization of freshwater prawn of genus Macrobrachium in the coastal area of Cameroon.

Macrobrachium (Bate, 1868) is a large and cosmopolitan crustacean genus of high economic importance worldwide. We investigated the morphological and molecular identification of freshwater prawns of the genus Macrobrachium in South, South West, and Littoral regions of Cameroon. A total of 1,566 specimens were examined morphologically using a key described by Konan (Diversité morphologique et génétique des crevettes des genres Atya Leach, 1816 et Macrobrachium Bate, 1868 de Côte d'Ivoire, 2009, Université d'Abobo Adjamé, Côte d'Ivoire), leading to the identification of seven species of Macrobrachium: M. vollenhovenii (Herklots, 1857); M. macrobrachion (Herklots, 1851); M. sollaudii (De Man, 1912); M. dux (Lenz, 1910); M. chevalieri (Roux, 1935); M. felicinum (Holthuis, 1949); and an undescribed Macrobrachium species M. sp. To validate the genetic basis of the identified species, 94 individuals representing the species were selected and subjected to genetic characterization using 1,814 DArT markers. The admixture analysis revealed four groups: M. vollenhovenii and M. macrobrachion; M. chevalieri; M. felicinum and M. sp; and M. dux and M. sollaudii. But, the principal component analysis (PCA) separated M. sp and M. felicinum to create additional group (i.e., five groups). Based on these findings, M. vollenhovenii and M. macrobrachion may be conspecific, as well as M. dux and M. sollaudii, while M. felicinum and M. sp seems to be different species, suggesting a potential conflict between the morphological identification key and the genetic basis underlying speciation and species allocation for Macrobrachium. These results are valuable in informing breeding design and genetic resource conservation programs for Macrobrachium in Africa.

Milk Composition for Admixed Dairy Cattle in Tanzania.

It is well established that milk composition is affected by the breed and genotype of a cow. The present study investigated the relationship between the proportion of exotic genes and milk composition in Tanzanian crossbred dairy cows. Milk samples were collected from 209 animals kept under smallholder production systems in Rungwe and Lushoto districts of Tanzania. The milk samples were analyzed for the content of components including fat, protein, casein, lactose, solids-not-fat (SNF), and the total solids (TS) through infrared spectroscopy using Milko-Scan FT1 analyzer (Foss Electric, Denmark). Hair samples for DNA analysis were collected from individual cows and breed composition determined using 150,000 single nucleotide polymorphism (SNP) markers. Cows were grouped into four genetic classes based on the proportion of exotic genes present: 25-49, 50-74, 75-84, and >84%, to mimic a backcross to indigenous zebu breed, F1, F2, and F3 crosses, respectively. The breed types were defined based on international commercial dairy breeds as follows: RG (Norwegian Red X Friesian, Norwegian Red X Guernsey, and Norwegian Red X Jersey crosses); RH (Norwegian Red X Holstein crosses); RZ (Norwegian Red X Zebu and Norwegian Red X N'Dama crosses); and ZR (Zebu X GIR, Zebu X Norwegian Red, and Zebu X Holstein crosses). Results obtained indicate low variation in milk composition traits between genetic groups and breed types. For all the milk traits except milk total protein and casein content, no significant differences (p < 0.05) were observed among genetic groups. Protein content was significantly (p < 0.05) higher for genetic group 75-84% at 3.4 ± 0.08% compared to 3.18 ± 0.07% for genetic group >84%. Casein content was significantly lower for genetic group >84% (2.98 ± 0.05%) compared to 3.18 ± 0.09 and 3.16 ± 0.06% for genetic group 25-49 and 75-84%, respectively (p < 0.05). There was no significant difference (p < 0.05) between breed types with respect to milk composition traits. These results suggest that selection of breed types to be used in smallholder systems need not pay much emphasis on milk quality differences as most admixed animals would have similar milk composition profiles. However, a larger sample size would be required to quantify any meaningful differences between groups.

Signatures of Selection in Admixed Dairy Cattle in Tanzania.

Multiple studies have investigated selection signatures in domestic cattle and other species. However, there is a dearth of information about the response to selection in genomes of highly admixed crossbred cattle in relation to production and adaptation to tropical environments. In this study, we evaluated 839 admixed crossbred cows sampled from two major dairy regions in Tanzania namely Rungwe and Lushoto districts, in order to understand their genetic architecture and detect genomic regions showing preferential selection. Animals were genotyped at 150,000 SNP loci using the Geneseek Genomic Profiler (GGP) High Density (HD) SNP array. Population structure analysis showed a large within-population genetic diversity in the study animals with a high degree of variation in admixture ranging between 7 and 100% taurine genes (dairyness) of mostly Holstein and Friesian ancestry. We explored evidence of selection signatures using three statistical methods (iHS, XP-EHH, and pcadapt). Selection signature analysis identified 108 candidate selection regions in the study population. Annotation of these regions yielded interesting genes potentially under strong positive selection including ABCG2, ABCC2, XKR4, LYN, TGS1, TOX, HERC6, KIT, PLAG1, CHCHD7, NCAPG, and LCORL that are involved in multiple biological pathways underlying production and adaptation processes. Several candidate selection regions showed an excess of African taurine ancestral allele dosage. Our results provide further useful insight into potential selective sweeps in the genome of admixed cattle with possible adaptive and productive importance. Further investigations will be necessary to better characterize these candidate regions with respect to their functional significance to tropical adaptations for dairy cattle.

Use of High Density Single Nucleotide Polymorphism (SNP) Arrays to Assess Genetic Diversity and Population Structure of Dairy Cattle in Smallholder Dairy Systems: The Case of Girinka Programme in Rwanda.

In most smallholder dairy programmes, farmers are not fully benefitting from the genetic potential of their dairy cows. This is in part due to the mismatch between the available genotypes and the environment, including management, in which the animals perform. With sparse performance and pedigree records in smallholder dairy farms, the true degree of baseline genetic variability and breed composition is not known and hence rendering any genetic improvement initiative difficult to implement. Using the Girinka programme of Rwanda as an exemplar, the current study was aimed at better understanding the genetic diversity and population structure of dairy cattle in the smallholder dairy farm set up. Further, the association between farmer self-reported cow genotypes and genetically determined genotypes was investigated. The average heterozygosity estimates were highest (0.38 ± 0.13) for Rwandan dairy cattle and lowest for Gir and N'Dama (0.18 ± 0.19 and 0.25 ± 0.20, respectively). Systematic characterization of the genetic variation and diversity available may inform the formulation of sustainable improvement strategies such as targeting and matching the genotype of cows to productivity goals and farmer profile and hence reducing the negative impact of genotype by environment interaction.

Computational genome-wide identification of heat shock protein genes in the bovine genome.

Background: Heat shock proteins (HSPs) are molecular chaperones known to bind and sequester client proteins under stress. Methods: To identify and better understand some of these proteins, we carried out a computational genome-wide survey of the bovine genome. For this, HSP sequences from each subfamily (sHSP, HSP40, HSP70 and HSP90) were used to search the Pfam (Protein family) database, for identifying exact HSP domain sequences based on the hidden Markov model. ProtParam tool was used to compute potential physico-chemical parameters detectable from a protein sequence. Evolutionary trace (ET) method was used to extract evolutionarily functional residues of a homologous protein family. Results: We computationally identified 67 genes made up of 10, 43, 10 and 4 genes belonging to small HSP, HSP40, HSP70 and HSP90 families respectively. These genes were widely dispersed across the bovine genome, except in chromosomes 24, 26 and 27, which lack bovine HSP genes. We found an uncharacterized outer dense fiber ( ODF1) gene in cattle with an intact alpha crystallin domain, like other small HSPs. Physico-chemical characteristic of aliphatic index was higher in HSP70 and HSP90 gene families, compared to small HSP and HSP40. Grand average hydropathy showed that small HSP (sHSP), HSP40, HSP70 and HSP90 genes had negative values except for DNAJC22, a member of HSP40 gene family. The uniqueness of DNAJA3 and DNAJB13 among HSP40 members, based on multiple sequence alignment, evolutionary trace analysis and sequence identity dendrograms, suggests evolutionary distinct structural and functional features, with unique roles in substrate recognition and chaperone functions. The monophyletic pattern of the sequence identity dendrograms of cattle, human and mouse HSP sequences suggests functional similarities. Conclusions: Our computational results demonstrate the first-pass in-silico identification of heat shock proteins and calls for further investigation to better understand their functional roles and mechanisms in Bovidae.