-
Welcome to BirdForum, the internet's largest birding community with thousands of members from all over the world. The forums are dedicated to wild birds, birding, binoculars and equipment and all that goes with it.
Please register for an account to take part in the discussions in the forum, post your pictures in the gallery and more.
What's next? (1 Viewer)
- Thread starter gusasp
- Start date
Acanthis
Well-known member
The Crithagra (former Serinus) finches of Africa would benefit from a thorough study. It's very likely there's more than one genus involved there.
Plus there's those odd species like Serinus nigripes and Ankober Serin (Carduelis/Serinus/Crithagra?) that definitely need a closer look
Plus there's those odd species like Serinus nigripes and Ankober Serin (Carduelis/Serinus/Crithagra?) that definitely need a closer look
l_raty
laurent raty
Available data:
GenBank: currently one ODC sequence.
BOLD: nothing.
The Genbank sequence is from Jønsson et al. 2011 [pdf incl. supporting info].
In this paper, Androphobus clusters with very high support (BS>95, PP>0.99 in Fig. S1) with Psophodes olivaceus. This is the expected position of the taxon based on morphology: so far, so good.
There is a potential problem here, though. Look at the data set (Table S1, lineage #31): Androphobus is represented by a single sequence of ODC; Psophodes is represented by sequences of myoglobin, ND2, and RAG1. Thus there is not a single base pair of DNA in the data set that is homologous between these two taxa. This is quite amazing, because this means that it is actually perfectly impossible, with this data set, to compute a divergence measure between these two taxa, or to identify any synapomorphy that would be shared by them. Were we to modify the trees of Fig. S1, by moving the node uniting them backwards or upwards (be it to the point that it merges with the directly more inclusive node in a trichotomy), this would have absolutely no effect on the goodness of fit of the tree...
This node can now be checked directly, though, as an ODC sequence of Psophodes olivaceus was recently made available [GenBank] by Aggerbeck et al. 2014 [pdf]. I have joined a tree based on selected ODC core-corvoid sequences: in this tree, Androphobus is sister to Psophodes with a 96% BS. Androphobus and Psophodes also share an apparently synapomorphic 11bp insertion in their ODC sequence, that reinforces their relationship.
l_raty
laurent raty
Available data:
GenBank: currently two partial cox1 sequence.
BOLD: nothing.
The cox1 sequences are from Sonet et al. 2001 [pdf].
One is 100bp, the other almost 300bp, which is a bit less than a half classical barcode.
100bp is really short; with the 297bp seq, I'd expect that it should be possible to link the species to reasonably a close relative. The closest other barcodes currently available are roughly 12% away from Graueria, and at this distance, the signal is clearly lost. OTOH, no potential relatives of Gaueria have been barcoded up to now, so maybe one day...
l_raty
laurent raty
Available data:
GenBank: one partial cyt-b sequence of G. lunullata.
BOLD: nothing.
The cyt-b sequence was obtained by Groenenberg, Davison, Chahil, & Schilthuizen, and is linked to an unpublished re-analysis of the relationships among peacock-pheasants. This sequence being six years old, though, it seems quite likely that the study will ever appear in press. This sequence was used by Stein 2013 [pdf], in an attempt to construct a general, multi-gene phylochronology of the Galliformes. Stein's conclusion was that it was not sufficient to give a stable position to the taxon, and that it destabilized his trees, so he finally excluded it from his analyzes. Note that a quite significant number of Galliform cyt-b sequences in GenBank show signs of problems (misidentifications, contaminations, chimeras); I did not attempt to replicate Stein's data set exactly but, on a quick scan, it seems to have included some sequences that I would not have used.
I have joined a tree based on selected Galliform cyt-b sequences: in this tree, Galloperdix clusters with Polyplectron and Haematortyx (Polyplectronini sensu TiF). Support is limited, though.
l_raty
laurent raty
Available data:
GenBank: one αA haemoglobin subunit sequence.
BOLD: nothing.
The haemoglobin sequence is from McCracken et al. 2009 [pdf] (I presume that it may have been used to build the tree presented in Fig.2 of this paper, but this paper was about haemoglobin evolution in high-altitude species, and the species that did not fit the subject are not labelled in the tree).
I have joined a tree based on selected Anseriform hba2 sequences: in this tree, Plectopterus gambensis is sister to Pteronetta hartlaubi with quite good support (86% BS), within a well-supported Anatinae (87.5% BS).
Note also the paraphyly of Chloephaga (visible with many genes), polyphyly of Netta and, on the contrary, well-supported monophyly of the traditional Anas.
Richard Klim
-------------------------
Laurent, impressive work for an otherwise quiet Sunday afternoon!
Peter Kovalik
Well-known member
l_raty
laurent raty
Available data:
GenBank: two cox1 barcodes of R. eurizonoides from the Philippines; one complete mitochondrial genome (appearing two times) of R. eurizonoides sepiaria from Japan.
BOLD: 13 barcodes, three of which are public--one of R. fasciata, the other being mined from GenBank, and actually representing the two instances of the Japanese complete mitochondrial genome.
The complete mt genome is from Ozaki et al. 2010 [pdf]: see Fig. 2 in this paper: the species is clearly embedded in Rallidae. The Philippine barcodes that appear in GenBank were recently deposited by Luczon et al., and are part of an unpublished barcoding study of the birds of the Philippines; they are also in BOLD (there are links to BOLD in GenBank), but are not yet public there.
See also the multigene tree in my post here.
Available data:
GenBank: 20 sequences of various nuclear markers.
BOLD: nothing.
Most Genbank data are from Hackett et al. 2008 [pdf] (there is also a sequence from Yuri et al. 2008 [pdf], and one from Braun et al. 2011 [pdf]). From Hackett et al. 2008, Himantornis is clearly a Rallidae, not at all close to Psophiidae.
See also the βFib7 tree in my post here, where I combined data from Hackett et al with other Gruiform βFib7 found in GenBank: in this tree, Himantornis doesn't even appear to be particularly basal among rallids. It seems closer to the Rallus-Gallirallus clade than Fulica, Gallinula, Aramides, Laterallus, Coturnicops or Porphyrio.
l_raty
laurent raty
It's only a small part of the list, Richard... (Don't know if I'll really go through the whole thing, though... We'll see.
)
andrew147
Well-known member
l_raty
laurent raty
Lots of data, long time to bootstrap...
Available data:
GenBank: 19 sequences of various genes.
BOLD: 2 public barcodes.
Among the GenBank sequences are 13 nuclear marker sequences from Hackett et al. 2008 [pdf], one from Yuri et al. 2008 [pdf] and one from Braun et al. 2011 [pdf]), as well as three mitochondrial sequences (12S, cyt-b, ND2) and a RAG1 sequence from Pereira & Baker 2010 [pdf]. Hackett et al. 2008 did not include any Glareolidae; in their data set Dromas was sister to Larus, and these two sister to Turnix. Yuri et al. 2008 and Braun et al. 2011 do not help to answer the question either. Pereira & Baker 2010 found Dromas closer to Glareolidae, but support was low and this result was lost when they combined their molecular data with morphological data.
Among Hackett et al.'s markers, two have also been sequenced for members of Glareolidae: myo (myoglobin) (sequences by Ericson et al. 2003 [pdf]) and βFib7 (sequence by Fain & Houde 2004 [pdf]). These two genes, when analyzed in isolation, suggested Dromas closer to Glareolidae -- with myo, the support was low (in a 200-replicate analysis, I got 60.5% BS); with βFib7, it was strong (in a 200-replicate analysis, I got 94.5% BS). RAG1 alone grouped Dromas and Glareolidae with 75% BS; mtDNA (cox1 + cytb + 12S rRNA + ND2) produced a trichotomy (Glareolidae; Dromas; all other Lari).
I have joined a tree resulting from a multigene (cox1 sequences, completed with 12S, cytb, ND2, RAG1, myo and βFib7 sequences as available in GenBank; data set partitioned by gene, and by position for coding genes [15 partitions in total; 3 for each coding gene (cox1, cytb, ND2, RAG1), 1 for the rRNA, 1 for each intron (myo, βFib)]) analysis of Lari, with Turnix and five Scolopaci as outgroups. There I recover Dromas as sister to Glareolidae with 87.5% BS, so my bet will go for that.
Note also that Procelsterna is nested in Anous (strong support, but the former is only represented by a cox1 sequence), and the very strong support given to (Larus + Chroicocephalus sensu BOURC TSC) (= Larus sensu AERC TAC), with the Saunders's Gull basal to this clade.
l_raty
laurent raty
l_raty
laurent raty
Available data:
Ptilorrhoa in GenBank: 21 sequences.
Ptilorrhoa in BOLD: 4 barcodes but none public.
Cinclosoma in GenBank: 161 sequences.
Cinclosoma in BOLD: nothing.
The genes that have sequences from both genera in GenBank are as follows:
cytb
Ptilorrhoa: 1 sequence from Norman et al. 2009 [pdf]
Cinclosoma: 1 sequence from Norman et al. 2009 [pdf]
ND2
Ptilorrhoa: 5 sequences from Toon et al. 2012 [abstract]
Cinclosoma: 33 sequences from Toon et al. 2012 [abstract]
myo
Ptilorrhoa: 1 sequence from Jønsson et al. 2007 [pdf], 1 sequence from Norman et al. 2009 [pdf]
Cinclosoma: 1 sequence from Norman et al. 2009 [pdf], 1 sequence from Aggerbeck et al. 2013 [pdf]
ODC
Ptilorrhoa: 1 sequence from Jønsson et al. 2007 [pdf]
Cinclosoma: 1 sequence from Aggerbeck et al. 2013 [pdf]
RAG1
Ptilorrhoa: 1 sequence from Barker et al. 2004 [pdf], 1 sequence from Norman et al. 2009 [pdf]
Cinclosoma: 1 sequence from Norman et al. 2009 [pdf]
GAPDH
Ptilorrhoa: 1 sequence from Jønsson et al. 2007 [pdf], 1 sequence from Toon et al. 2012 [abstract]
Cinclosoma: 16 sequences from Toon et al. 2012 [abstract], 1 sequence from Aggerbeck et al. 2013 [pdf]
Among the works that produced these sequences only two sampled both taxa: Norman et al. 2009 and Toon et al. 2012. Toon et al. 2012 addressed specifically the evolution of Cinclosoma, using Ptilorrhoa sequences as outgroup in their analyzes, and sampled no other taxa: this paper assumed, but does not allow to assess the relationships of the two genera. Norman et al. 2009 used cytb, myo and RAG1; they published three single-gene trees and a tree resulting from a combined analysis of the three genes. Cinclosoma and Ptilorrhoa were sister in all these trees (albeit support was lacking with cytb).
I obtained similar results with the other three genes: Cinclosoma and Ptilorrhoa were sister with strong support based on ODC and ND2; GAPDH also placed them together, but this didn't survive bootstrapping (too short sequences). Thus I don't actually see any suggestion of disagreement in the data. Did I miss something?
Below, I have attached a tree derived from a combined analysis of ND2, cytb, myo, ODC, RAG1, GAPDH, TGFβ2, brahma, and βFib5. This is a bit more than what is available for the Ptilorrhoa/Cinclosoma pair: I did this to be able to use the same analysis to assess some of the other cases. In this tree, Cinclosoma and Ptilorrhoa are sister with 100% BS.
(Note - If anyone wants to see single-gene trees, please, ask.)
Last edited:
andrew147
Well-known member
Aggerbeck et al. (2014) have Cinclosoma in Pachycephalidae, as sister to Falcunculus. This seems bizarre and implausible on the surface, but stranger things have happened...
http://www.sciencedirect.com/science/article/pii/S1055790313003850
Consequently, TiF has moved Cinclosoma to the Pachycephalidae but retained Ptilorrhoa in Psophodidae - ergo, presumed closest relatives end up in separate families.
Last edited:
l_raty
laurent raty
OK. The data I've looked at do not resolve the position of the pair, and the position of Falcunculus is undetermined as well. But Aggerbeck et al. had 11 genes for Cinclosoma, and I used only 5 of these here. (The 6 others have not been sequenced for Ptilorrhoa.)
But given the support there is for the Cinclosoma-Ptilorrhoa relationship, I think that if Cinclosoma is moved somewhere, Ptilorrhoa should follow.
But given the support there is for the Cinclosoma-Ptilorrhoa relationship, I think that if Cinclosoma is moved somewhere, Ptilorrhoa should follow.
Last edited:
Peter Kovalik
Well-known member
It seems the Ploceidae family is a hodge-podge of genera, e.g. Anaplectes here and other in Jetz et al.
Excerpt from Jetz et al Genetic data only trees attached.
