From: Robert Hazen Sent: Thursday, March 10, 2016 22:10 To: Downs, Robert T - (rdowns) Cc: Daniel Hummer; Jesse H. Ausubel; Golden, Joshua J - (jgolden); Edward Grew; Chao Liu; Morrison, Shaunna - (shaunnamm); Grethe Hystad; Sergey Krivovichev; Peter Fox; Marshall X Ma; Michael Meyer; Paul Falkowski; Dimitri Sverjensky; Jihua Hao; Elisha K. Moore Subject: Re: Paper idea--mineral rarity biosognatures Attachments: HAZEN_2013_AJS.pdf Flag Status: Flagged Dear Colleagues (including Keck colleagues not copied on the previous messages), We have begun a conversation asking if the distribution of rare minerals on Earth is a biosignature. I am now proposing to use the Keck data-science workshop at RPI (June 8-10) as a forum to learn how to mine these data. There is a significant amount of advanced database work we'll need before that, but we have some concrete questions in mind. Please read the previous messages for context. First, Jesse, Dan and Bob, thanks for various thoughts. Very useful feedback. I agree with Jesse that we should do all mineral species. I agree with Dan that considering contrasting distributions on Moon and Mars is an important challenge, though I wonder if we yet have enough data for definitive answers, but we should look carefully at Grethe's work on Mars mineralogy. I have a suggestion for one way to proceed with the question of distributions of paragenetic modes among commoner versus rarer minerals, and whether we can tease out a biosognature based on mineral diversity/distribution. A while ago I did a survey of all the Handbook of Mineralogy (HoM) and tabulated a few dozen different paragenetic modes for minerals; some are obviously biological and some not. I didn't make any tables at that time except for the 420 presumed Hadean minerals (article attached; check out table 1, last column). Paragenetic mode is an important parameter in characterizing mineral species. So, could we add a field to the rruff.info/ima database in which we have a code to designate all the known paragenetic modes for each of the ~5100 mineral species? We could divide up this chore; I'd do a lot of it. I would first send around a proposed list of all mineral paragenetic modes with a key--a number from 1 to 40, for example, for each distinct mode. I'd make that list flexible, in case we want to revise and/or expand the list. Then several of us would start to populate the database. The objective would be to have all data completed before the June 8-10 workshop. We could then use the mindat (or other) lists of numbers of known localities for each species, and do some kind of graph analysis that merged many mineral parameters: composition, age, rarity, paragenetic mode, crystal structure complexity, tectonic setting, and whatever else seemed interesting. Then let Peter's group show us ways to analyze these multi-dimensional data for 5100 mineral species to see if we have strong correlations with age or mode of formation (especially biotic vs. abiotic) or some combination of parameters. My gut feeling is that, if our databases are sufficiently populated and we ask the right questions, we could generate several papers from what we find in June. So, action items: 1. I'll draft a list of paragenetic modes with some kind of key; distribute and ask for comments. 2. I'll ask Josh and Bob D. how we might best add a field in the rruff database for paragenetic mode, and how we might devise an easy protocol for all of us to begin the one-time (potentially tedious) process of entering 5100 codes. I assume that we'd all have a spreadsheet that lists all mineral species and we'd just add a column with the paragenetic mode numbers. Then Josh could automatically transfer the data to the new field. 3. We divide up the work in some way for those who are willing to join in. I'll do a lot of it; it's really easy to scan through the Handbook of Mineralogy (hard copy is easiest, arranged alphabetically, flip page-by-page from the 6 volumes; just add numbers to the spreadsheet Josh will provide). That will get us about 75% of the minerals. The rest will probably require a combination of geochemical intuition (i.e., for oxidized weathered phases like hydrous sulfates and arsenates--easy to spot and invariably post-GOE), and looking at the primary literature. Shoot to finish prior to June 8. Once and done. I volunteer to start with the first 3 volumes of HoM. Anyone else? Could be pretty interesting. Best, Bob On Thu, Mar 10, 2016 at 11:48 AM, Downs, Robert T - (rdowns) wrote: Recall that in Grethe’s last paper which is still in review, she shows that while we have only identified about 30 minerals on Mars, their distributions is consistent with LNRE. Grethe might be able to state this better than I. Bob From: Daniel Hummer [mailto:dhummer@carnegiescience.edu] Sent: Thursday, March 10, 2016 8:49 AM To: Jesse H. Ausubel ; 'Robert Hazen' Cc: Downs, Robert T - (rdowns) ; Golden, Joshua J - (jgolden) ; Edward Grew ; Daniel Hummer ; Chao Liu ; Morrison, Shaunna - (shaunnamm) ; Grethe Hystad Subject: Re: Paper idea--mineral rarity biosognatures Bob and folks, This is a thoroughly fascinating idea. I can envision it being a very daunting task to characterize the paragenetic modes of all 5100 minerals currently in the catalogue. But I can also see Jesse's point, because if we're talking about doing 2800, we're already over halfway to the full list, and there may be value in saying that we did them all. What would be really fascinating is to compare the mineral distribution curves of Earth to that of Mars and the Moon, but I doubt that enough "localities" on the Mars and Moon have been catalogued to do meaningful statistics. But if we do the full list of terrestrial minerals, and compare the distribution of biotic vs. abiotic vs. all minerals, we might find interesting differences. Even if they are all LNRE distributions, the parameters of the distributions might differ in some significant way that constitutes a biosignature. I did something similar to this when I compared element correlations among all minerals with element correlations among Hadean minerals at AGU (from a list Bob had already developed from a previous paper). It turns out the main difference was that a living Earth contained a dramatic abundance of O-H correlations, undoubtedly because of the proliferation of biologically-mediated hydrous minerals. It may be that a similar kind of biosignature exists in the very nature of the distributions..... maybe looking at the locality distribution among minerals in the Hadean minerals list from Bob's AJS paper (which we know have major paragenetic modes that are abiotic) is a good place to start - maybe this list shows a lower abundance of "rare" minerals than a LNRE distribution would predict? What does everyone think? If we settle on a list that's too large, we can always divide labor between several people to make it more manageable. What I'm describing is perhaps another way of getting at the same question, but I also like Bob's idea of checking for a clear difference in paragenetic mode among rare vs. common species. I think either approach will yield similar information. Dan Daniel R Hummer Postdoctoral Scholar, Mineralogy and Crystallography Geophysical Laboratory Carnegie Institution of Washington Phone: (814) 321-8859 Fax: (202) 478-8901 On 3/10/16 10:23 AM, Jesse H. Ausubel wrote: Dear Bob, This is very interesting but I am unsure why one would restrict the search to the 1800 + 1000. It might be tedious, but why not do the entire 5500? Characterizing and perhaps classifying the entire set of “biominerals” would seem valuable and deeper. Best- Jesse From: Robert Hazen [mailto:rhazen@ciw.edu] Sent: Thursday, March 10, 2016 10:18 AM To: Jesse H. Ausubel; Bob Downs; Joshua Golden; Edward Grew; Daniel Hummer; Chao Liu; Shaunna Morrison; Grethe Hystad Subject: Paper idea--mineral rarity biosognatures Dear Colleagues, In Paris. Just back from the second of two seminars on mineral evolution/ecology, the first yesterday at the Sorbonne, today at IPGP. Very positive feedback. At both seminars, early-career people asked about the relationship between LNRE distributions and biosignatures--something to which we allude in a couple of papers (but hadn't mentioned explicitly in my talks, so these were pretty perceptive questions). Anyway, walking back from IPGP I realized how we might easily test the hypothesis that the LNRE distribution, particularly the distribution of the rarest minerals, might reflect biological mediation. A corollary to the hypothesis (as yet untestable) would be that Mars and the Moon do not display LNRE mineral distributions. So, we have about 1800 species found at only 1 or 2 localities (that's the un- massaged mindat.org data). It's pretty easy to check most of those species to see if they're biologically derived. Anything that's an oxidized weathering product, anything that's an organic mineral, and species that arise exclusively through human processes would be included. Based on a recent partial survey, my hunch is that many of the remaining non-biological rare species will prove to be rare because they are difficult to collect--(e.g., form in volcanic fumaroles), not because they are inherently rare (see the recent Hazen & Ausubel paper, attached). So those minerals could be much more common than indicated by collections. So, if we do a survey of the 1800 rarest species we should be able to compare their paragenetic modes to those of the 1000 most common species and see if the distributions of paragenetic modes differ. Looking at paragenetic modes of 2800 species is a bit of a slog, but made much easier by the Handbook of Mineralogy hard copy, which is arranged alphabetically so one can just flip through the pages. All I need at present is for Josh to send me a spread sheet of the ~1800 minerals from only 1 or 2 mindat localities arranged alphabetically, and a list of the 1000 commonest minerals (in terms of mindat localities). Then I'll check them all and compare the two sets. If anyone else is excited by this and wants to chip in, fine. Otherwise I'll just do a few at a time and tabulate the distribution of paragenetic modes. Could be a pretty high-impact paper if we can show unambiguously that rare minerals constitute a biosignature. I welcome thoughts. Will return to DC on Sunday, and will keep you posted. All the best, Bob -- Robert M. Hazen Senior Staff Scientist, Geophysical Laboratory Executive Director, Deep Carbon Observatory 5251 Broad Branch Road NW Washington, DC 20015 phone: 202-478-8962 e-mail: rhazen@ciw.edu Personal web site: http://hazen.gl.ciw.edu DCO website: deepcarbon.net Keck Deep-Time Project website: http://dtdi.carnegiescience.edu