Cordova and Sanford: Good Enough For A Secular Journal?

Continuing the discussion from Developing College-Level ID/Creation Courses:

I would like an honest and respectful assessment of the science in this paper. It is known that the authors are YEC. That is beside the point. Valid critique will not be directed at the author’s personal view themselves or “what they might do” if this paper was published. My question, rather, is concerning the scientific content in this paper alone.

Are their “conclusions right?”

What do the scholars think?

@art, @John_Harshman, @glipsnort, @nlents, @davecarlson and @evograd, what do you think?

1 Like

Thank you. Presently I’m not able to comment on that thread, and that’s fine. However, there are two figures missing from the pre-print, partly because the journal had us put the figures in another file and I didn’t get around to putting it in the pre-print arxiv, but just the text on the arxiv. Anyway, here are the figures:

Figure 1:

Figure 2:

Referring to Table 1, why only 1 psi-BLAST iteration? Why not just use BLASTP? How did you choose an e-value cutoff (or Expect threshold) of 10? (IMO, with this value, you are going to get many, many false positives.)

The preprint you shared, @stcordova, reads like a hit piece on Ohno. It is pretty rude and unprofessional. That is enough to reject it from most journals and other scientific venues.


The reason is even with this relaxed value we could not get Ohno’s PR.C to even register a positive. If there were many false positives for a beta lactamase, why not even a single positive (true or false) for one for PR.C!

We could have run an more constricted search, but given a relaxed search could not even return one positive hit for Ohno’s PR.C, then why would we expect a more constricted search to help. That was the point.

Fair enough, but was Ohno wrong.

Please show us some offending quotes @art.

Ohno didn’t have the same information as us. That paper was from decades ago. He does not deserve any disrespect, even if data from decades later shows him wrong.

Agree, he deserves respect. We can always re-write to be nicer, and such feedback is helpful. But was Ohno wrong?

1 Like

I’m sure, like all of us, Ohno was wrong about some things and right about other things. He was almost certainly right about frameshifts as a mechanism for de novo proteins.

The salient question for this paper, it seems, is whether or not nylonase B arose by a frameshift or not. Is that what you are asking?

Yes. We were specifically talking Nylonase Nyl B. We mentioned instances of other frameshifts that had reasonable evidence by Okamura.

Please provide those quotes too.

From the paper:

Overlapping reading frames are known to exist in biology, [26,27] and Okamura [28] has speculated that several such human genes may have originated via frame shift mutations. However, it is worth noting that Okamura believed that Ohno’s claims were correct when he proposed this. With the falsification of Ohno’s famous hypothesis, it becomes very reasonable to question the evolutionary importance of novel and useful proteins arising instantly by frameshift mutations.

  1. Rancurel C, Khosravi M, Dunker AK, Romero PR, Karlin D. Overlapping genes produce proteins with unusual sequence properties and offer insight into de novo protein creation. J Virol. 2009 Oct;83(20):10719-36. Doi: 10.1128/JVI.00595-09.
  1. Sherr CJ. Divorcing ARF and p53: an unsettled case. Nat Rev Cancer. 2006 Sep;6(9):663-73.
  1. Okamura K, Feuk L, Marques-Bonet T, Navarro A, Scherer SW. Frequent appearance of novel protein sequences by frameshift translation. Genomics. 2006 December;88(6):600-697.

I believe some stuff can come from frameshift mutation.

NOTE: the reason I had a presentation a few months ago on promiscuous domains is that proteins with promiscuous domains are the sort of proteins that are NOT expected to arise via frame-shift mutations unless its some sort of back mutation reactivating a gene. But the issue of promiscuous domains wasn’t mentioned in our paper.

My slides again on promiscuous domains:

I don’t see how this conclusion follows at all. Did all the other researchers who’ve identified putative proteins arising from frameshift mutations make the same mistakes that Ohno apparently did?


I agree @davecarlson. It is possible that the nylonase B example was falsified. This does not falsify the Ohno hypothesis that frameshifts are important in the Genesis of many de novo proteins. The prototypical example could be false, while the hypothesis survives due to other validated examples.

On a technical level, I’m also suspicious of the Psi Blast with only one iteration. Why? That doesn’t make sense.


Is in contrast with:

Where in the paper do you note that the evidence for these other examples is reasonable?

The wording in the paper could perhaps be re-written, but this was the original

An objection might be raised that the psi-BLAST and SPARCLE results were artificially inflated by redundancies and spurious hits for NylB because of the search parameters we used. This objection would not be valid. First, these parameters were the defaults set by the NIH, and second, the point of the comparison was to show that even under relaxed and generous parameters, no remote potential homologs of PR.C could be detected

spurious hits are false positives. But the point was to give the most relaxed possible seach for PR.C and not NylB. And even then PR.C couldn’t register hits. NylB’s abundance and other nylonases was well demonstrated by the UniProt searches and didn’t require us to do a BLAST search at all. Establishing the homology of NylB to beta-lactamases was quite sufficient to establish the ubiquity of NylB and its homologs. The psi-BLAST was not done for the benefit of demonstraing NylB ubiquity, but rather PR.C’s abundance, and even under the most relaxed parameters, it didn’t return a single hit. Therefore it didn’t exist, and therefore the frame-shift hypothesis for NylB is wrong and should stop being cited.

I thought referencing Okamura’s paper showed it was reasonable. I contacted Dr. Okamura and he generously responded. SInce his publication, there have been a few issues with removal from RefSeq the proteins he used as examples. But since this whole thing was outside the scope of the paper, we didn’t pursue it more.

But frameshift mutations can’t make important domains like the ones I’ve discussed, therefore, the bar for “de novo” seems awfully low. But that’s outside the scope of the paper.

Okay. That is all the more reason to use more psi BLAST iterations.

It looks like you were searching using amino acids in a blastp based search, is that right?

To answer this question, you would need to do the searches with the DNA sequence in a blastn search. Did you do a blastn search?

If the frameshift hypothesis is true, we’d expect no hit with blastp (perhaps your current results) with a clear hit with blastn.

1 Like