User talk:Fylindfotberserk/Archives/2010/March

Sharma 2009

Your argument, that the error was typographical, is not borne out by the facts.

• It is clear from the paper that Sharma et al. claim the "72.22%" number as their own finding. (And they do so in a revised version of their paper: that is, they had ample opportunity to check for errors.)

• From the abstract, top of p.47: "A peculiar observation of the highest frequency (up to 72.22%) of Y haplogroup R1a1* in Brahmins hinted at its presence as a founder lineage for this caste group."
• Top of p.51b: "Further, the observation of a very high frequency (upto 72.22%) in this study (Table 1) [...]"

• It is clear that Sharma 2009 claim to have screened 30 West Bengal Brahmins.

• The number 30 appears in Table 1 in a column that correctly totals to 621, the very number for total sample size stated in the abstract and also at the top of p.50b ("Genetic structure of the studied regional population groups").

IOW, if you are correct about the real source of the 72.22% number, then this is much much more than a matter of erroneous entries in a table of data. Given your hypothesis, Sharma 2009 have claimed someone else's finding as their own: they have "borrowed" the Sengupta 2006 numbers for 18 WB Brahmins and projected them on their own 30 WB Brahmins. Do you know what this would imply about their scientific integrity? You may have missed that the case of West Bengal Brahmins is not the only problem with the data in Sharma 2009. See User:Rudrasharman/Notes/DNA papers#Sharma 2009 this. 8 out of 16 rows are thoroughly compromised. How many other sources did they plunder for their claimed sample of 621? (And note also, in the paper, at p.54a, they claim to have found 19/71 Saharia males with R1a1*. The reported sample size in Table 1, consistent with the reported total sample size of 621, was 57. Is that a "typographical error" too?)
Would you still want to claim that Sharma 2009 has any shred of scientific value whatsoever? If so, on what basis?
(Please respond here, to keep discussion in one place. I'll check back. Thanks.) rudra (talk) 17:55, 1 March 2010 (UTC)

Sorry for the late reply…
If you look at my posts, I definitely did not deny that Sharma et al guys might have claimed the WB Brahmin data of 72.22% as their own, instead I focused on the fact that the percentage is correct for that specific group and should not have been removed from the wiki article just because of it being highlighted in Sharma et al. Actually I was quite amazed to see the 72.22% listing referenced from Sharma et al 2009 in an older version of this article sometime in Jan-Feb 2009 as I’d already calculated that same Y-DNA percentage from Sengupta et al. But, instead of making a direct allegation towards them, I said that they might have ‘compiled’ their own data with other sources which actually is not an offence as most of the geneticists use data from previous studies to further their own analysis unless they did not mention it explicitly and in this case they probably did not(except for the Maharashtra Tribes in the Supplementary Doc.).
Besides, my ‘hypothesis’ that ‘Sharma et al has claimed Sengupta data’ is not only borne out of the PDF but from the attached ‘Supplementary Doc’ as well. Here they used the data of West Bengal Brahmins and other Indian, Pakistani, Central Asian groups that they likely have taken from other studies and compiled alongwith their own data of 510 subjects (not 621) for MDS plots shown in Figure 3C and Figure 5 of the doc. It proves that WB Brahmin, Himachal Brahmin and other groups were ‘referenced’ in their own study which included only 510 samples of the groups listed here.
Now, as far as me quoting the samples of WB Brahmins as typo error is concerned, ‘18 samples projected as 30 samples’ looks like a typo error for that group and does have the potential to inflate the no. from 609 to 621. And the same might have occurred for other compromised groups like Punjab Brahmins, Himachal Brahmins, Saharia, etc. However if they had already picked up a total no. of samples of 621 before analysis as you’re pointing to, then they probably inflated the referenced data by big margins. They also made minor adjustments to their own data set of MP Gonds and Maharashtra Brahmins in the published PDF as I’ve referenced here.
Saharia data is seriously messed up and this group doesn't show up in the supplementary. MP Tribe in the doc refers to MP Gonds only. --Fylindfotberserk (talk) 16:39, 8 March 2010 (UTC)