KELVIN: A Software Package for Rigorous Measurement of Statistical Evidence in Human Genetics

Vieland, Veronica J.; Huang, Yungui; Seok, Sang-Cheol; Burian, John; Çatalyürek, Ümit V.; O’Connell, Jeffrey R.; Segre, Alberto M.; Valentine-Cooper, William

doi:10.1159/000330634

Cited by 34 publications

(50 citation statements)

References 102 publications

(77 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Here we focus only on the EqS itself and not its derivation. This EqS turned out to be a function of two aspects of the LR (and two constants; see below): (i) the logarithm of the maximum LR, which we treated as an entropy term (see Appendix A) and denoted as S; and (ii) the area under the LR graph, denoted V, which is related to, though distinct from, the Bayes factor [10] and the Bayes ratio in statistical genetics [11]. Originally these equations were given as:…”

Section: Review Of Previous Results and The Problem With Two-sided Hymentioning

confidence: 99%

Statistical Evidence Measured on a Properly Calibrated Scale across Nested and Non-nested Hypothesis Comparisons

Vieland

Seok

2015

Entropy

Self Cite

View full text Add to dashboard Cite

Statistical modeling is often used to measure the strength of evidence for or against hypotheses about given data. We have previously proposed an information-dynamic framework in support of a properly calibrated measurement scale for statistical evidence, borrowing some mathematics from thermodynamics, and showing how an evidential analogue of the ideal gas equation of state could be used to measure evidence for a one-sided binomial hypothesis comparison ("coin is fair" vs. "coin is biased towards heads"). Here we take three important steps forward in generalizing the framework beyond this simple example, albeit still in the context of the binomial model. We: (1) extend the scope of application to other forms of hypothesis comparison; (2) show that doing so requires only the original ideal gas equation plus one simple extension, which has the form of the Van der Waals equation; (3) begin to develop the principles required to resolve a key constant, which enables us to calibrate the measurement scale across applications, and which we find to be related to the familiar statistical concept of degrees of freedom. This paper thus moves our information-dynamic theory substantially closer to the goal of producing a practical, properly calibrated measure of statistical evidence for use in general applications.

show abstract

Section: Review Of Previous Results and The Problem With Two-sided Hymentioning

confidence: 99%

Statistical Evidence Measured on a Properly Calibrated Scale across Nested and Non-nested Hypothesis Comparisons

Vieland

Seok

2015

Entropy

Self Cite

View full text Add to dashboard Cite

show abstract

“…We conducted a genome-wide family-based linkage analysis using the Bayesian PPL method [25] on 3 families that had at least one prodigy and at least one separate person with ASD and 2 prodigy-only families. To assess if prodigy and autism have shared loci, we coded both prodigy and autism as the same trait status, a model that therefore implies etiological equivalence for the two diagnostic categories despite differing presentations.…”

Section: Resultsmentioning

confidence: 99%

“…We chose KELVIN [25] to conduct primary linkage and association analyses. KELVIN implements a Bayesian analysis framework that performs linkage, LD (also known as association analysis), and joint linkage-LD analysis, all within the same framework.…”

Section: Linkage/association Analysis Methodsmentioning

confidence: 99%

Molecular Genetic Evidence for Shared Etiology of Autism and Prodigy

Ruthsatz¹,

Petrill²,

et al. 2015

Hum Hered

View full text Add to dashboard Cite

Child prodigies are rare individuals with an exceptional working memory and unique attentional skills that may facilitate the attainment of professional skill levels at an age well before what is observed in the general population. Some characteristics of prodigy have been observed to be quantitatively similar to those observed in autism spectrum disorder (ASD), suggesting possible shared etiology, though objectively validated prodigies are so rare that evidence has been sparse. We performed a family-based genome-wide linkage analysis on 5 nuclear and extended families to search for genetic loci that influence the presence of both prodigy and ASD, assuming that the two traits have the same genetic etiology in the analysis model in order to find shared loci. A shared locus on chromosome 1p31-q21 reached genome-wide significance with two extended family-based linkage methods consisting of the Bayesian PPL method and the LOD score maximized over the trait parameters (i.e., MOD), yielding a simulation-based empirical significance of p = 0.000742 and p = 0.000133, respectively. Within linkage regions, we performed association analysis and assessed if copy number variants could account for the linkage signal. No evidence of specificity for either the prodigy or the ASD trait was observed. This finding suggests that a locus on chromosome 1 increases the likelihood of both prodigy and autism in these families.

show abstract

“…Hence we have log 2 units of information for each observed toss. Then if we know the actual sequence of tosses, we have I SEQ = log 2 n = n log 2 (4) units of information for the full set of n tosses. (For present purposes the base of the logarithm, which further defines the units of information, is unimportant.…”

Section: The 2nd Law Of Thermodynamics and Information Lossmentioning

confidence: 99%

“…presents itself naturally. Previously in this journal [1][2][3][4] and elsewhere [5,6] , my colleagues and I have advocated the need for an absolute, context-independent scale for measurement of evidence in statistical genetics and other fields. We have illustrated situations in which the lack of such a measure leads to erroneous inference, e.g., the conclusion that an original linkage or association report was a false positive when in fact subsequent data support the original finding.…”

Section: Introductionmentioning

confidence: 99%

Evidence, Temperature, and the Laws of Thermodynamics

Vieland

2014

Hum Hered

Self Cite

View full text Add to dashboard Cite

A primary purpose of statistical analysis in genetics is the measurement of the strength of evidence for or against hypotheses. As with any type of measurement, a properly calibrated measurement scale is necessary if we want to be able to meaningfully compare degrees of evidence across genetic data sets, across different types of genetic studies and/or across distinct experimental modalities. In previous papers in this journal and elsewhere, my colleagues and I have argued that geneticists ought to care about the scale on which statistical evidence is measured, and we have proposed the Kelvin temperature scale as a template for a context-independent measurement scale for statistical evidence. Moreover, we have claimed that, mathematically speaking, evidence and temperature may be one and the same thing. On first blush, this might seem absurd. Temperature is a property of systems following certain laws of nature (in particular, the 1st and 2nd Law of Thermodynamics) involving very physical quantities (e.g., energy) and processes (e.g., mechanical work). But what do the laws of thermodynamics have to do with statistical systems? Here I address that question.

show abstract

KELVIN: A Software Package for Rigorous Measurement of Statistical Evidence in Human Genetics

Cited by 34 publications

References 102 publications

Statistical Evidence Measured on a Properly Calibrated Scale across Nested and Non-nested Hypothesis Comparisons

Statistical Evidence Measured on a Properly Calibrated Scale across Nested and Non-nested Hypothesis Comparisons

Molecular Genetic Evidence for Shared Etiology of Autism and Prodigy

Evidence, Temperature, and the Laws of Thermodynamics

Contact Info

Product

Resources

About