Attribution, Prediction, and the Causal Interpretation ...

Attribution, Prediction, and the Causal Interpretation ...

d e t i s i v e R y t i l i b a

t i r e H be u c n M e Z i nhl What is heritability? Heritability estimates have been regarded as important primarily on the expectation that they would furnish valuable information about

the causal strength of genetic influence on phenotypic differences (Sesardic, 1993:399) h2 (in the broad sense) = proportion of total phenotypic variance in a population (VP ) attributable to genetic variance (VG) h 2 = VG / VP E.g. Height differs among people If height has a high h2 in a population, then differences in height in that population are mostly due to genetic differences in the population rather than any environmental differences Why should we care about h2? h2 is widely used in human disease genetics, amongst other fields, for e.g. Tenesa and Haley (2013:147) explain: Estimates of heritability quantify how much of the variation in disease liability in a population can be attributed to genetic variation

Vissher et al. (2008:258-9) argue that h2 is so enduring and useful because, amongst other reasons, it is useful in understanding the genetic component of risk to disease, independently of known environmental risk factors. More specifically, it useful in determining the efficiency of prediction of the genetic risk of disease (ibid.). Heritability today Despite this kind of use of heritability estimates, widely held view among philosophers today = heritability estimates do not indicate the causal strength of genes on phenotypic variance (Oftedal, 2005; Downes, 2016) Specifically 3 main lines of arguments are used to undermine the causal construal of some heritability claims: 1. gene-environment interaction 2. gene-environment correlation

3. locality On heritability In this talk I argue that the widely held view that heritability estimates are devoid of causal implications (Sesardic, 2005:10) is too quick It is possible to reply to each of the main lines of argument used to establish that heritability estimates are causally uninterpretable In particular, these 3 challenges to heritability analysis express conditions under which heritability claims can be well-justified, possibly generalizable, depending on empirical matters that cannot be spelt out in advance The gene-environment challenge There are in fact 2 different notions of gene-environment

interaction (Tabery, 2007a; 2007b; 2008, 2014; Griffiths and Tabery, 2008), each with different problems for causal construal of h2 First notion = biometric notion of gene-environment interaction, GEB The concept of GE [GEB] refers to cases where different genotypic groups phenotypically respond differently to the same array of environments. (Tabery, 2007a:2) (From Kaplan, 2006:60, Figure 1) when different genetic groups respond differently to the same array of environments, the additivity between VG and VE breaks down,

requiring an addition to the equation in the form of GE [GEB]. GE creates a potential problem for biometricians because it generates its own variation [GEB] ... eliminating the ability to calculate the heritability of a trait (Tabery, 2007a:120). The GEB challenge (B1): h2 analysis is based on the assumption of additivity (differences in genotype and differences in environments act separately on total phenotypic variance); (B2): GEB complicates the calculation of h2 because it generates a separate source of variation in the total phenotypic variance sum (Tabery, 2007a); (B3): If there is strong GEB, then heritability cannot be easily calculated for a particular trait (ibid.); (B4): Non-additivity is rampant in nature (e.g. Lewontin, 1974, 1975; Layzer, 1974; Block & Dworkin, 1974b; Dick & Rose, 2002); GEB is

pervasive in nature; (B5): Therefore, heritability estimates cannot have a causal interpretation, let alone a useful one (Lewontin, 1974; Northcott, 2008:22); or heritability is meaningless in terms of its causal explanatory content (Tal, 2011) Reply to the GxEB challenge A) The additivity reply: the question of additivity, of the actual pervasiveness of GEB in nature, is an empirical one Heritability estimates can have a causal interpretation when there is no statistical gene-environment interaction (amongst other conditions). The GEB challenge turns on the pervasiveness of GEB in nature. But if the question of non-additivity is an empirical one (one that cannot be decided a priori), if significant GEB is rare, then the GEB

can be dissolved The question we now have to consider is, is it true that non-additivity is rampant in nature? Reply to the GEB challenge Like Jensen (1969), Levin (1997), Sesardic (2005), Oftedal (2005), and Tal (2009, 2011), I agree that we cannot answer the question of the pervasiveness of GEB and non-additivity in biology in a nonempirical way To the effect that step 4 (B4) in the GEB challenge to heritability (non-additivity is rampant in nature) is defended in an aprioristic manner (for e.g. by Gray (1999) & Sterelny and Griffiths (1999)), it cannot be used to establish the strong conclusion that heritability estimates have no causal import If GEB is substantial in a particular study, then calculating heritability will be futile (Tal, 2009, 2011) - uncontroversial But when there is no GEB , it makes sense to causally interpret the

heritability estimate (Sesardic, 2005; Tal, 2011) Lessons on heritability Possible problems? 1) At exactly which point or which percentage (of total phenotypic variance) does statistical gene-environment interaction become problematic for heritability? 2) Methodological issue that the analysis of variance cannot detect some statistical geneenvironment interaction effects Challenges to heritability analysis express conditions

under which heritability claims can be well-justified, possibly generalizable, depending on empirical matters that cannot be spelt out in advance Once we deal with the question, does it ever make sense to causally interpret heritability claims? We can now ask, what nonstatistical reality underlies heritability claims? Heritability Revisited 1.What is heritability? 2.Why should we care about heritability? 3.The gene-environment challenge to heritability analysis

4.Lessons on heritability u o y k Than

Recently Viewed Presentations

  • The Challenge: To Create More Value in All Negotiations

    The Challenge: To Create More Value in All Negotiations

    need MicrosoftWord FONTS "Showcard Gothic," "Ravie," "Chiller" and "Verdana" Master* Excellence part one (of 3) introduction to excellence.
  • Home Hemodialysis: Patient Selection and the Importance of

    Home Hemodialysis: Patient Selection and the Importance of

    Informed Choice Obstacles to Home Hemodialysis. Historically, up to 88% of patients in the US reported being unaware that HHD was an option. Lack of modality education typically leads to in-center HD as the default modality, particularly for late-referred patients
  • Mental Health and Wellbeing Thursday 3rd December 2015

    Mental Health and Wellbeing Thursday 3rd December 2015

    Dörte Bemme is a PhD candidate at McGill University. Her research involves multi-sited fieldwork within the institutional assemblage of Global Mental Health, investigating how GMH re-configures the knowledges and practices that define what a 'mental health' problem is and how...
  • Defense Property Accountability System Presented by Mr. Joseph

    Defense Property Accountability System Presented by Mr. Joseph

    Print Work Orders. Track Warranties, Services and Subscriptions for assets. Record Asset Utilization in Miles/Hours/Shot Fired. Fuel Type and Quantity Consumed. Dispatch Capabilities. Recurring / Ad Hoc. Reserve by Dispatch Sub Category or Specific Item. Equipment Pool. Generate Dispatch/Utilization Form....
  • Social Security www.socialsecurity.gov History of Social Security Programs

    Social Security www.socialsecurity.gov History of Social Security Programs

    This also applies to an appointed representative or someone who is in business with the number holder. You may be unable to open a . my. ... Once you give us this additional information, we will send you a letter...
  • ABPI POI Project Recommendations

    ABPI POI Project Recommendations

    O'Neil Report 2014 (final report 2016) Slide . Impact of AMR on World Economy. ... Need to maintain momentum, we have a fragile consensus, at a tipping point. Collaboration. Industry is ready and willing - with Government and all stakeholders.
  • Earned Media - Kiwanis

    Earned Media - Kiwanis

    -Vicki. Look at the subject line for the email: Man hosts Christmas party for needy children for past 35 years. That's interesting for two reasons - it's a Christmas party for needy children and he's been hosting it for 35...
  • WELCOME MELBOURNE ALUMNI TO OUR SEMINAR Superfoods or

    WELCOME MELBOURNE ALUMNI TO OUR SEMINAR Superfoods or

    Cherry-sized purple berry fruit of the acai palm. Lab studies suggest it may have anti-cancer and anti-inflammatory effects, as well as a possible use in treating heart disease