<%--FREQUENTLY ASKED QUESTIONS FOR BISGOAT
--%>
1. Can I compare my one breed sample with all recognised goat breed of India?
Ans.Yes, the BIS-G server database contains 55000 allelic data which is from 22 recognized registered goat breeds of India viz. Blackbengal, Ganjam, Gohilwari, Jharkhandblack, Attapaddy, Changthangi, Kutchi, Mehsana, Sirohi, Malabari, Jamunapari, Jhakarana, Surti, Gaddi, Marwari, Barbari, Beetal, Kanniadu, Sangamnari, Osmanabadi, Zalawari and Cheghu. Each goat breed in the database is represented by 50 genetically unrelated goat animals having true to the breed type phenotype from native breeding tract of each goat breed.
They work as standard reference data in computational analysis for breed prediction.
2. If I have admixture, can I get quantified degree of admixture?
Ans. Yes, if your goat query is with an unknown or a crossbred or breed with potentially different degree of admixture, you will get degree of admixture with quantified value in a pie chart.
3. From where can I get my goat breeds genotype data (DNA data)for breed test using this server?
Ans. Your blood samples/carcass samples/biological samples can be sent to any lab with microsatellite genotyping facility or to any commercial vendors along with primer sequence given on the website or its link to get data.
4. What is the probability of getting wrongly predicted goat breed?
Ans. The DNA based prediction gives very high accuracy and in our method, it has been up to more than 95 %, however if the sample belongs to a new breed which has not been used in the training data, that may not give as high accuracy as mentioned.
5. If genotype value for any locus is missing, then what would be done?
Ans. If genotype value for any locus is missing, one should put the value of zero on server while uploading the data. Such missing values of one or two locus usually do not compromise with accuracy of breed prediction.
6. Being dinucleotide repeat as most predominant in large animal/advance eukaryotes is it necessary to have constancy of odd or even numbers of base pairs in STR allele? Whether such experimental mistake (Type II error) will affect the computational part?
Ans.Odd or even number of base pair is often confused but in reality it may vary. For example even if locus repeats are even in numbers the flanking region (including primer) is never fixed thus it will vary from odd to even. Beside this the locus itself may be simple or compound or compound interrupted repeat. Thus this misconception should not be considered as mistake of -------------- result potentially affecting accuracy of prediction.
We have given range of base pair size expected in mazor Indian goat population as cross check for fully accuracy of genotyping.
GLOSSARY
Activation function: A function used to transform the activation level of a unit (neuron) into an output signal. Typically, activation functions have a "squashing" effect. Together with the PSP function (which is applied first), this defines the unit type. Neural Networks supports a wide range of activation functions.
Only a few of these are used by default; the others are available for customization.
Allele: Any of two or more genes/DNA sequences that have the same relative position on homologous chromosomes and are responsible for alternative characteristics
Artificial Neural Networks: These are analytic techniques modelled after the (hypothesized) processes of learning in the cognitive system and the neurological functions of the brain and capable of predicting new observations (on specific variables) from other observations (on the same or other variables) after executing a process of so-called learning from existing data.
Breed: A group of organisms having common ancestors and certain distinguishable characteristics
Cross entropy: Error functions based on information-theoretic measures, and particularly appropriate for classification networks. There are two versions, for single-output networks and multiple-output networks; these should be combined with the logistic and softmax activation functions respectively.
Distribution: An order or pattern formed by the tendency of a sufficiently large number of observations to group themselves around a central value.
DNA marker: Alleles of DNA polymorphisms, used as experimental tags to keep track of an individual, a tissue, a cell, a nucleus, a chromosome, or a gene. Stated another way, any character that acts as a signpost or signal of the presence OR location of a gene OR heredity characteristic in an individual of a population.
Error function: The error function is used in training the network and in reporting the error. The error function used can have a profound effect on the performance of training algorithms.
Feedforward neural networks: Neural networks with a distinct layered structure, with all connections feeding forwards from inputs towards outputs. Sometimes used as a synonym for multilayer perceptrons.
Genotype: The actual alleles present in an individual; the genetic makeup of an organism.
Hidden layer: All layers of a neural network except the input and output layers. Hidden layers provide the network's non-linear modeling capabilities.
Input layer: First layer of a ANN consists of input units. These units are known as independent variables in statistical literature.
Learning algorithm: The optimal weights may be obtained by using Gradient descent algorithm (GDA), Broyden-Fletcher-Goldfarb-Shanno (BFGS), or Conjugate gradient descent algorithm (CGDA) learning algorithms.
Learning rate: A control parameter of some training algorithms, which controls the step size when weights are iteratively adjusted.
Locus: The position on a chromosome of a gene or a particular segment of DNA (marker).
Microsatellite: Tandem repeats of short simple DNA sequence, generally of 1-6 bases.
Multilayer perceptron: Feedforward neural networks having linear PSP functions and (usually) non-linear activation functions.
Neuron: A unit in an artificial neural network.
Output layer: Last layer contains output units. In statistical nomenclature, these units are known as dependent or response variables.
Parameters: Parameters are used to identify a characteristic, a feature, a measurable factor that can help in defining a particular system
PCR: A method of DNA analysis that exponentially amplifies a specific DNA sequence or region allowing rapid DNA analysis
Population: A population is all the organisms that both belong to the same group or species and live in the same geographical area.
Post Synaptic Potential (PSP) Function: The function applied by a unit to its inputs, weights and thresholds to form the unit's input (or activation) level. The two major PSP functions are linear (weighted sum minus threshold) and radial (scaled squared distance of weight vector from input vector).
Primer: A short oligonucleotide sequence used to amplify DNA sequences in a polymerase chain reaction.
Radial basis function: A type of neural network employing a hidden layer of radial units and an output layer of linear units, and characterized by reasonably fast training and reasonably compact networks. Introduced by Broomhead and Lowe (1988) and Moody and Darkin (1989), they are described in most good neural network text books.
STATISTICA Automated Neural Networks: STATISTICA Automated Neural Networks (SANN) contains a comprehensive array of statistics, charting options, network architectures, and training algorithms; C and PMML (Predictive Model Markup Language) code generators. The C code generator is an add-on.
Supervised learning: The term "supervised" learning is usually applied to cases in which a particular classification is already observed and recorded in a training sample, and want to build a model to predict those classifications (in a new testing sample). The purpose of the classification analysis would be to build a model to predict who (from a different list of new potential customers) is likely to respond to the same (or a similar) offer in the future. These methods are called supervised learning algorithms because the learning (fitting of models) is "guided" or "supervised" by the observed classifications recorded in the data file.
Unsupervised learning: In unsupervised learning, the outcome variable of interest is not (and perhaps cannot be) directly observed. Instead, we want to detect some "structure" or clusters in the data that may not be trivially observable.