Citation
Van Geystelen A., Decorte R., Larmuseau M.H.D. (2013),
Updating the Y-chromosomal phylogenetic tree for forensic applications
based on whole genome SNPs
Forensic Science International: Genetics, 7 (6), 573-580.
Archived version
Author manuscript: the content is identical to the content of the published
paper, but without the final typesetting by the publisher
Published version
insert
link
to
the
published
version
of
your
paper
http://www.sciencedirect.com/science/article/pii/S1872497313000896
Journal homepage
insert
link
to
the
journal
homepage
of
your
http://www.fsigenetics.com/home.
Author contact
your email maarten.larmuseau@bio.kuleuven.be
Klik hier als u tekst wilt invoeren.
IR
url in Lirias https://lirias.kuleuven.be/handle/123456789/ 395607
(article begins on next page)
paper
Abstract
The Y-chromosomal phylogenetic tree has a wide variety of important forensic
applications and therefore it needs to be state-of-the-art. Nevertheless, since the
last 'official' published tree many publications reported additional Ychromosomal lineages and other phylogenetic topologies. Therefore, it is
difficult for forensic scientists to interpret those reports and use an up-to-date
tree and corresponding nomenclature in their daily work. Whole genome
sequencing (WGS) data is useful to verify and optimize the current
phylogenetic tree for haploid markers. The AMY-tree software is the first open
access program which analyses WGS data for Y-chromosomal phylogenetic
applications. Here, all published information is collected in a phylogenetic tree
and the correctness of this tree is checked based on the first large analysis of
747 WGS samples with AMY-tree. The obtained result is one phylogenetic tree
with all peer-reviewed reported Y-SNPs without the observed recurrent and
ambiguous mutations. Nevertheless, the results showed that currently only the
genomes of a limited set of Y-chromosomal (sub-)haplogroups is available and
that many newly reported Y-SNPs based on WGS projects are false positives,
even with high sequencing coverage methods. This study demonstrates the
usefulness of AMY-tree in the process of checking the quality of the present Ychromosomal tree and it accentuates the difficulties to enlarge this tree based on
only WGS methods.
Keywords
Haploid markers, Y-chromosome, phylogenetic tree, bio-informatics, Wholegenome SNP calling, Y-SNPs
Introduction
A state-of-the-art phylogeny of the human Y-chromosome based on bi-allelic
polymorphisms is an essential tool for forensic genetics. Forensic scientists are
taking advantage of the Y-chromosomal phylogenetic tree in their daily work,
e.g. by checking the quality of datasets or by assigning geographical landscapes
to specific lineages [1, 2]. Y-chromosomal single nucleotide polymorphisms
(Y-SNPs) have a great capacity for detecting geographical origins as many
lineages defined by Y-SNPs show a strong continent-specific [3, 4] and even
intra-continent-specific distribution [5-7]. Their usefulness is illustrated by the
fact that Y-SNP data are now also included in Y-chromosomal forensic
databases, such as in the YHRD database [8]. Therefore, an up-to-date extended
Y-chromosomal phylogeny based on these bi-allelic markers which are
preferably unambiguous and non-recurrent but which have a high
discrimination power is required for forensic applications.
Since the publication of the latest 'official' Y-chromosomal phylogenetic tree by
Karafet et al. [4], a continuous wave of new peer-reviewed articles which report
changes to this tree are published. These changes include a new root and new
basal clades [9, 10], modifications of the global backbone [3, 11], different
phylogenetic topologies within a haplogroup [12-14], newly described subhaplogroups [15-17], or other phylogenetic positions for a certain mutation
[18]. As these publications are not coordinated different names are given to the
Y-chromosomal lineages for which the phylogenetic position is given in
different topologies. Therefore, the currently overall reported Y-chromosomal
tree is not clear and this makes it difficult for forensic researchers to use a
uniform phylogenetic tree. Hence new initiatives to ensure more continuity in
the report of the most recent phylogenetic Y-chromosomal tree are needed.
Large whole genome sequencing (WGS) projects such as the 1000 Genomes
project [19, 20] bring an opportunity to introduce the required uniformity in the
reporting of the haploid Y-chromosomal tree. The analysis of whole Ychromosomes within male genomes allows verification and optimization of the
currently used phylogenetic tree. WGS data has already proved to be useful in
verifying and optimizing the phylogeny of the other haploid markers in the
human genome i.e. the mitochondrial DNA (mtDNA). Relevant ambiguous
markers and back-mutations which influence the interpretation of previous
forensic and evolutionary genetic studies were detected based on these data
[21]. Recently a new Y-chromosomal phylogenetic tree was built after a tabula
rasa of the present Y-chromosomal tree by using only Y-SNPs from available
WGS male samples [22]. By comparing this new phylogenetic tree with the
currently used one, the backbone of the currently used phylogenetic tree was
confirmed in this study. However, this new tree is not useful for forensic
research because there is no link between currently used and newly reported
lineages. Furthermore, the set of used genomes is not a good representation of
all existing Y-chromosomal (sub-)haplogroups and geographical regions. There
are also still too much false positive SNP calls in this WGS dataset.
Alternatively, the AMY-tree software is the first open access program which
academics and forensic professionals can use to verify and optimize the
currently used Y-chromosomal tree by using WGS data [23]. The first AMYtree analysis was done based on 118 WGS samples and proved already its
usefulness to verify and to optimize the present Y-chromosomal phylogenetic
tree [23].
The aim of this study is to perform the largest reported screening of male
genomes for Y-chromosomal phylogenetic applications based on the AMY-tree
software. Firstly, we want to merge all newly Y-SNPs from recent peerreviewed publications since the latest 'official' Y-chromosomal phylogeny [4]
into one single tree which is useful for forensic applications. Secondly, this
updated Y-chromosomal phylogenetic tree needs to be checked for recurrent
mutations, ambiguous SNPs and other difficulties for the (forensic) application
of the tree. Thirdly, this study also wants to find out for which Y-chromosomal
(sub-)haplogroups there is already WGS data available. Finally, investigating
the possibilities to enlarge the Y-chromosomal phylogenetic tree based on the
current Y-SNP detections in WGS data is the last aim of this study.
Materials and methods
Updated phylogenetic tree
The latest updated phylogenetic tree of the Y-chromosome as it was published
by Van Geystelen et al. [23] was manually updated based on recent descriptions
of new Y-SNPs in academic research papers like Pamjav et al. [16] and
Scozzari et al. [9]. As the exact phylogenetic position of a few new Y-SNPs
was not given their position needed to be determined based on the results of
AMY-tree of all WGS samples. Next, also recurrent mutations, ambiguous
SNP-loci and wrongly defined mutation conversions within the newly updated
Y-chromosomal tree were ascertained based on those AMY-tree results.
WGS Y-SNPs dataset
In order to check the manually updated phylogenetic tree and to optimise the
AMY-tree software, a large dataset of whole genome Y-SNP calls was
assembled. This dataset consists of 747 samples which represent 660 males, as
several genomes were analysed in different projects. Within this dataset the
genomes of eight males whose father's genome was also sequenced are present.
The SNP calls were collected from four large WGS projects and several
individual genome projects (Supplementary Materials Table S1). These projects
differ from each other based on the used next-generation sequencing (NGS)
platforms and sequence coverage. First, Complete Genomics made the SNP
calls
of
35
whole
genomes
of
males
available
(http://www.completegenomics.com/public-data/69-Genomes/ 04 Jan 2013);
those genomes were sequenced with a high sequencing coverage on the
Complete Genomics Analysis (CGA) Platform [24]. Second, the Personal
Genome Project (PGP) and Singapore Sequencing Malay Project (SSMP) also
used this CGA platform. PGP is a project started to obtain and openly share
human genome sequences in combination with health information. At the
moment 40 male genomes were available (www.personalgenomes.org 04 Jan
2013). The SSMP on the other hand wanted to characterize the polymorphic
variants in the population of Malays, an Austronesian group present in
Southeast Asia and Oceania. Recently, the Y-SNP calls of 46 Malays were
made publically available [25]. Next, the 1000 Genomes Project aims to
provide a comprehensive resource on human genetic variation by sequencing
more than 1000 human genomes. In 2010, SNP calls of 77 males were made
available in the pilot phase [20] and two years later a set of 526 SNPs profiles
were published as result of phase 1 of the Project [19]. As the 1000 Genomes
project aims to sequence a large number of people, the sequencing coverage
was lower than in the other projects. Finally, 23 additional samples were
collected from several single genome projects [26-35, 36 and unpublished
genomes of Guy Froyen].
AMY-tree modifications
Several modifications to the AMY-tree software version 1.0 [23] were made for
the assessment of the SNP calling quality in WGS data. This was necessary as
the quality of the SNP calling influences the AMY-tree analysis of a sample
and therefore also the interpretation of the result of the analysis [23]. The extra
quality assessment is based on the results of the first AMY-tree run of a certain
sample. This assessment assumes that the used phylogenetic tree is correct and
that the assigned haplogroup after the first AMY-tree run is the actual
haplogroup of that sample.
The algorithm for the extra quality assessment is simple and comprehensible as
shown in Figure 1. First, all Y-SNPs of the phylogenetic tree are selected
except if the determined haplogroup of the first run is a paragroup (e.g.
R1b1b2*). In the case of a paragroup, all Y-SNPs which are in sub-nodes of the
main group of this paragroup are excluded from the selection. This is done in
order to remove the influence of too much false positive SNP calls because
when the haplogroup is not at the correct phylogenetic level, too much false
positives and/or false negatives mutant Y-SNPs would be detected. Next, the
expected state of all selected Y-SNPs is determined whereby all Y-SNPs in the
path from the assigned sub-haplogroup till root are expected to be mutant and
all others are expected to be ancestral. The state of each Y-SNP is also
determined in the reference genome. Thereafter, these expectations of Y-SNP
state are converted into expectations of ‘called’ or ‘not called’ in the next step
based on the expected state and the state in the reference genome. Consider a
Y-SNP in the path from sub-haplogroup till root, if the status of that SNP in the
reference sequence is mutant, the Y-SNP is expected to be 'not called' else the
expectation is 'called'. Consider a Y-SNP not in the path from sub-haplogroup
till root, if the status of that Y-SNP in the reference sequence is mutant, the YSNP is expected to be 'called' else the expectation is 'not called'. Then, these
expectations are compared to the state in the sample such that the number of
true positive, false positive, true negative and false negative SNP calls will be
determined. At last, the quality of a sample is expressed in several measures of
quality: Matthews correlation coefficient (MCC), accuracy, sensitivity,
specificity, precision, recall and F1-score (Supplementary method). When the
MCC is larger or equal to 0.95 the SNP calling quality is called excellent,
otherwise it is called low. The quality will be given in the output file of AMYtree. When the SNP calling quality is low, caution has to be taken about the
result of the AMY-tree analysis due to the high occurrence of false negative and
false positive SNPs. When the quality is excellent, the results of AMY-tree are
considered to be valuable for the control of the currently used phylogenetic tree
of the Y-chromosome and for the increase of its resolution. As such, a better YSNP call quality assessment is implemented in AMY-tree version 1.1 compared
to the earlier version [23].
Next, when a sample was run in AMY-tree in the 'sufficient' mode such that the
reference genome was taken into account but the determined haplogroup
belongs to R-M269 and the MCC lower than 0.95 a second AMY-tree run
needs to be executed but in the 'insufficient' mode. This is important as a MCC
lower than 0.95 indicates that this result of the first run is too much influenced
by the reference genome. Finally, another small modification to AMY-tree is
based on the fact that Z381, L2 as well as L20 are mutant in the reference
genome and although AMY-tree version 1.0 already had a quite complex
system to filter out the influence of the reference genome even on samples
belonging to haplogroup R, it was not yet efficient enough. Therefore, when
both Z381 and L2 or both Z381 and L20 are mutant in the first AMY-tree run,
e.g. the ancestral SNP was not called, the sample will be handled as insufficient
in a second AMY-tree run, such that the Y-SNPs of the reference genome are
not used anymore when determining the haplogroup. By including these
modifications even more certainty is build in AMY-tree version 1.1 in
comparison with the previous version [23]: for samples belonging to a R-M269
sub-haplogroup the reference genome is only taken into account when the SNP
calling quality is excellent after the first run in the ‘sufficient’ mode.
The cut-offs to assess the Y-SNP calling quality is optimised based on all 747
genomes by performing several test runs and manual analyses of the genomes
and by checking it with the results in Van Geystelen et al. [23] and with the
publications of the genomes whereby a Y-chromosomal analysis was already
performed.
Y-SNP detecting
The Y-SNPs which are present in the WGS samples but which are not yet
included in the updated phylogenetic tree were detected by AMY-tree. Only
those Y-SNPs from WGS samples with an excellent Y-SNP calling quality
were used. But that was not the only constraint for the potentially relevant YSNPs; they must also be positioned in the unique regions of the Y-chromosome
as read mapping and variant detection difficulties are expected due to the high
frequency of repeated sequences on the Y-chromosome. So, Y-SNPs in the
pseudoautosomal, heterochromatic, X-transposed and ampliconic segments [37]
of the male-specific part of the genome as reported by Wei et al. [22] were
excluded.
Results
In total 747 samples are analysed by the updated version of the AMY-tree
software with the updated phylogenetic Y-chromosomal tree. There are 131
samples of 126 individuals with an excellent Y-SNP calling quality, i.e. MCC ≥
0.95, which are mostly obtained from Complete Genomics and the Personal
Genome Project. The remaining 616 samples have a low calling quality and are
mostly obtained from the 1000 Genomes Project pilot and phase 1 (Table S1).
Updated tree for forensic applications
The state-of-the-art Y-chromosomal tree is manually updated based on all
published Y-SNPs from academic studies. After the AMY-tree runs of all 747
samples, all Y-SNPs of which no exact phylogenetic position was given in the
publications could now be included in the updated phylogenetic tree, for
example there were two new Y-SNPs reported in [16] within sub-haplogroup
R-M198 without clear phylogenetic positions (Figure 2).
The results with an excellent SNP calling quality also showed several recurrent
Y-SNPs in the phylogenetic tree which cause the determination of multiple
haplogroups for some samples. After ruling out that the recurrent SNPs are
sample- or project-specific, these SNPs are removed from the phylogenetic tree;
an overview of the three observed recurrent SNPs of which enough evidence
was available, is given in Table S2. These modifications led to the final updated
tree version 1.1 which includes 359 Y-chromosomal lineages and 721 Y-SNP
markers. The final tree and its corresponding mutation conversions for all the
Y-SNPs in the tree can be found in Tables S3 and S4.
Sub-haplogroup determining
The determined haplogroups of all 747 samples obtained by the AMY-tree
analysis based on the final updated tree version 1.1 is given in Table S1. Only
for 14 samples no haplogroup can be determined and for three other samples
multiple haplogroups are determined by AMY-tree. The distribution of the
MCC of the 730 samples for which an unambiguous haplogroup is determined
is given in Figure 3. Only a minority of the samples has an excellent SNP
calling quality with a MCC ≥ 0.95; a MCC lower than 0.95 means that less than
97.5% of the negative and positive predictions are correct. Overall, 17 different
haplogroups and 106 sub-haplogroups are present in the dataset. When
considering only the samples of excellent quality 10 different haplogroups and
47 sub-haplogroups remain. Figure 4 and Table S5 give an overview of those
(sub-)haplogroups and their frequencies.
The samples of paternally related samples present in the dataset are of particular
interest because they are considered to represent the same Y-chromosome. All
the samples of one family with eight members sequenced by Complete
Genomics are determined to belong to R-P312*. The paternal grandfather of
that family is also analysed in the 1000 Genomes project and there he is
determined as P* [P-92R7*]. The first attempt of AMY-tree to determine the
sub-haplogroup of that sample of 1000 Genomes led to the sub-haplogroup
R-L2 which has a higher phylogenetic level than the haplogroups of the
Complete Genomics samples. However, as the MCC value of that sample is
smaller than 0.95 the sub-haplogroup determination is done again but without
the influence of the reference genome. This led to the final sub-haplogroup P92R7* which has a less accurate phylogenetic level than that of Complete
Genomics. Thus, the influence of the reference genome can sometimes cause a
too high or too low phylogenetic level when the new modifications which were
made to AMY-tree version 1.1 would not have been applied.
Detecting new Y-SNPs
The large amount of available samples leads to a huge amount of newly
reported Y-SNPs, i.e. Y-SNPs that are not yet present in the updated
phylogenetic tree version 1.1. In total 108,681 new Y-SNPs are reported in all
660 male genomes; when an individual is analysed in more than one project
only the sample with the highest MCC value is used. The majority of the SNPs
appears in only a few samples: 62% appears in only one sample and 16% is
present in two samples as shown in Figure 5A. In the 126 male genomes with
an excellent Y-SNP calling quality 50,430 new Y-SNPs are reported. These
SNPs also come with a high frequency of low occurrences in the excellent
genomes: 57% is unique and 11% appears in two samples. When only the
regions within the Y-chromosome which are identified as unique are taken into
account, a much lower number of new Y-SNPs is detected. In total 35,503 new
Y-SNPs are reported in the 660 male genomes and 15,208 new Y-SNPs in the
genomes with excellent Y-SNP calling quality. The same patterns of occurrence
for these Y-SNPs are observed as with the non-filtered Y-SNPs: the majority of
the new reported Y-SNPs appeared in only one or two samples of the WGS
dataset as shown in Figure 5B.
The genomes of eight males whose biological father's genome is also sequenced
are present in the dataset: one family of eight males including the father, the son
and the six grandsons next to one father-son pair. In the family of eight paternal
relatives 5,155 new SNPs are reported on the whole Y-chromosome and a large
number of these SNPs is found in only one of the eight individuals as shown in
Figure 6A. The number of Y-SNPs decreases every time the number of samples
in which the Y-SNP occurs increases except for the occurrence in all-exceptone and all samples of the family. When all Y-SNPs which also occur in any
other sample with excellent SNP calling quality are removed only less than
20% of the Y-SNPs remains but the distribution of the number of Y-SNPs per
occurrence stays the same as the black bars in Figure 6A show. The same
comparison is made when only the SNPs in the unique part of the Ychromosome are selected: the same pattern as with all SNPs is visible in Figure
6B. However, for each number of occurrences the number of truly unique
SNPs, i.e. SNPs that do not occur in other genomes outside the family, is much
higher. Within the father-son pair 2,181 new SNPs are reported on the whole Y-
chromosome for which the difference between occurrence in only one and both
genomes is relatively small, also for the proportion of SNPs which do not
occur in the other genomes with an excellent Y-SNP quality (Figure 6A).
Remarkably, there are more new SNPs present in both samples than in one
sample when the Y-SNPs which are not located in the unique region of the Ychromosome are removed (Figure 6B). The effect of the increasing number of
unique Y-SNPs that do not occur in other genomes as seen with the previous
family is also in the father-son pair present.
Discussion
The present study realises the first large screening of male genomes for
phylogenetic applications of the Y-chromosome. Based on this screening of 747
male samples an update has been made of the AMY-tree software and of a
state-of-the-art Y-chromosomal phylogenetic tree was established for forensic
scientists. Furthermore, also recommendations for future sequencing projects
dealing with a broader selection panel of Y-chromosomal haplogroup samples
and for the validation of newly detected Y-SNPs are made.
First, the large screening of the 747 male Y-chromosome samples revealed that
the SNP calling quality of a few samples was overestimated in AMY-tree
version 1.0. As these samples belong to sub-haplogroup R-M269 they are very
similar to the reference genome which is used to estimate the SNP calling
quality of a sample [23]. To remove this SNP calling quality overestimation the
influence of the reference genome is excluded for all samples belonging RM269 with a low SNP call quality. That is why several modifications are made
and implemented in AMY-tree version 1.1.
Second, an update of the currently used Y-chromosomal phylogenetic tree was
realised based on the large database of 747 available samples. At the moment
this tree is the most state-of-the-art tree applicable for forensic geneticists: all
Y-SNPs which are reported in academic publications till today are included and
all ambiguous markers are excluded to avoid wrong Y-SNP interpretations
(Table S3). As often the case in the literature the phylogenetic relationship
between new Y-SNPs and earlier reported Y-SNPs in the same (sub)haplogroup are not given. By using the AMY-tree results it is possible to find
out the concrete phylogenetic level of each Y-SNP relative to the other already
known SNPs. For example, Pamjav et al. [16] described and genotyped two
new Y-SNPs Z93 and Z280 within R-M198, although the exact phylogenetic
positions in relationship with the other lineages within R-M198 were not given.
The presence of both Z93 and Z80 is checked in all samples belonging to the
sub-haplogroup R-M198, Z93 occurred in three R-M198* samples and in none
of the other samples. Also Z280 does not occur in any sample except in one RM198* sample. Therefore, both Z93 and Z280 are placed in the phylogenetic
tree as sub-haplogroups of R-M198 as shown in Figure 2. The choice is made
for a phylogenetic tree in table format as described by Van Geystelen et al. [23]
instead of a branching diagram because the tree is very large and will become
larger in the future. Therefore, the table format can be adapted more easily than
the diagram and it is also more manageable.
Not only newly reported Y-SNPs are included in the new updated tree but also
ambiguous Y-SNPs are excluded as they can complicate the Y-chromosomal
applications for forensic studies. As previously described [38], the most
relevant ambiguous Y-SNPs are recurrent SNPs which have a paralogous
distribution along the phylogenetic tree and which have thus mutant alleles in at
least two independent Y-chromosomal lineages. Based on the screening of the
747 analysed samples, three Y-SNPs are recognized for the first time as
recurrent (Table S2). There are no other indications for recurrent mutations
based on the present WGS dataset. Therefore we may assume, in most cases,
that males which both have the mutant allele for a Y-SNP used in the updated
tree (Table S3) have received this mutant allele from one common ancestor and
not by convergent evolution. Next, also all Y-SNPs which could not be
analysed by WGS for one reason or another are excluded from the updated tree,
although it may be possible to genotype these SNPs correctly with Sanger
sequencing methods. For example, all hundred E-M2 samples in the dataset did
not reveal the mutant allele for Y-SNP V95 as expected by earlier publications
[14, 39]. The reason for this remarkable result may be that V95 is not well
validated but more likely it is the result of a bad SNP calling in the Ychromosomal region around V95 by the current WGS methods. For some YSNPs also a mutation conversion is found which is different from the reported
one. For example, another ancestral and mutant allele are observed for Y-SNP
M426 than reported by Rootsi et al. [40]. Since the Y-chromosome has a very
complex origin it also has a lot of non-unique regions which complicate the
analysis of WGS data [37]. The current reference genome GRCh37 shows the
evolution of the Y-chromosome and its numerous resulting non-unique regions.
So the reason for this wrong analysis of M426 may be the position of this SNP
in one of the non-unique regions of the Y-chromosome as defined earlier [22,
37]. To have an unambiguous phylogenetic tree, we excluded all the Y-SNPs
for which no reliable signal with the WGS methods could be found. In the end
an updated Y-chromosomal tree which includes 359 Y-chromosomal lineages
and more than 721 Y-SNP markers (Table S3) is obtained and this tree is the
basic tool to develop and optimize Y-SNP-multiplexes for forensic applications
[41, 42].
Third, the distribution of the analysed haplogroups per ancestral continent in the
current dataset corresponds with the known distributions [3, 4]. Nevertheless,
there is not yet a representative set of all phylogenetic sub-haplogroups
available. In total 17 haplogroups and 106 sub-haplogroups are reported in the
analysis of the whole dataset. However, the SNP calling quality is low (MCC <
0.95) for most of the analysed samples as most data are obtained from the 1000
Genomes project and therefore the sub-haplogroup determination of these data
is not completely reliable. When considering only the samples of excellent
quality (MCC ≥ 0.95) 10 haplogroups and 47 sub-haplogroups remain and this
corresponds with only 13% of the total number lineages described so far. Most
of the Y-chromosomes are assigned to haplogroups E, O and R. Therefore,
when the set of WGS Y-chromosomes will be enlarged the current phylogenetic
tree as well as the analysis of Wei et al. [22] to calibrate the tree can be
optimized.
Finally, the screening of the dataset also revealed that new in silico and in vitro
methods are required to verify new Y-SNPs based on WGS methods. As earlier
mentioned [23], a huge number of new Y-SNPs are false positives when the
genomes were sequenced with low coverage and consequently the called SNPs
will have a low quality. These false positive SNP calls disturb the determination
of the correct (sub-)haplogroup of the sample; consequently the AMY-tree
software has to correct for them by applying several additional methods in the
analysis [23]. The high number of false positive Y-SNPs - even detected in
WGS samples with an excellent SNP calling quality - is observed by comparing
genomes of paternal relatives. Within the eight paternally related samples
belonging to one family, which are genotyped by Complete Genomics, 949
newly reported Y-SNPs in the full Y-chromosome and 88 ones in the unique
regions of the Y-chromosome are found in at least one but not all family
members (Figure 6). Despite the high coverage of these genomes and the
excellent SNP calling quality this is still a very high number of newly reported
Y-SNPs which are most likely to be false based on the mutation rate on the Y-
chromosomes calculated based on a deep-rooting pedigree [43] and based on
human-chimpanzee comparisons [44]. A similar conclusion can be made based
on the father-son pair in the dataset which is also sequenced with a high
coverage by Complete Genomics (Figure 6).
The Y-SNP results show that adding new lineages to the Y-chromosomal
phylogenetic tree only based on WGS is not evident. Therefore, making tabula
rasa and building a new tree based on all WGS Y-chromosomes as done by
Wei et al. [22] is not an option when the tree is going to be used in forensics.
Each new Y-SNP and consequently each Y-chromosomal lineage has to be
validated independently from the WGS data before adding them to the updated
phylogenetic tree. The validation status of many potential polymorphic Y-SNPs
(> 1% in population) is often unclear but this can be resolved by the sharing of
genomic data among genetic genealogists which are interested in finding new
Y-SNPs to resolve their particular paternal ancestry [45]. Therefore, this is an
area in which closer collaborations between amateurs and forensic academics
could prove to be particularly useful [13]. Nevertheless, it is required that new
in silico methods will be designed to select good and relevant candidates of YSNPs for the validation. For example, an interesting criterion is the position of
the Y-SNPs: it is more interesting to validate only the SNPs located in the
unique regions of the Y-chromosome as there are many non-unique regions due
to the evolutionary history of this chromosome [37]. The lower number of false
positive SNP calls in these unique regions in comparison with those in the full
Y-chromosome is clearly demonstrated in the father-son pair. The number of
new Y-SNPs reported in only one of the two samples is higher than the new
ones reported in both samples based on the whole Y-chromosome (Figure 6A)
in contrast with the situation based on the unique regions of the Y-chromosome
(Figure 6B).
Conclusions
Based on the largest screening of male genomes with 747 samples in total, the
most up-to-date Y-chromosomal phylogenetic tree for forensic applications is
compiled. Future publications which will report new Y-SNPs have to situate
their phylogenetic positions in this tree to guarantee the continuity between old
and new publications. At this moment, forensic scientists as well as
evolutionary biologists and genetic genealogists are lost in the many reports of
newly described Y-chromosomal lineages [46]. Therefore, initiatives as AMYtree which optimize the phylogeny based on peer-reviewed publications are
required [23]. This is already the case for the mitochondrial genome with the
Phylotree initiative of van Oven and Kayser [47]. Nevertheless, to optimize the
current updated phylogenetic tree for the human Y-chromosome more high
quality genomes of a broader set of (sub-)haplogroups than the frequent
haplogroups E, O and R are required. Also a higher effort in the validation of
reported Y-SNPs by new in silico and in vitro methods is required.
Acknowledgements
The authors want to thank Tom Wenseleers, Manfred Kayser, Jean-Jacques
Cassiman, Tom Havenith, Hendrik Larmuseau and Lucrece Lernout for useful
discussions and comments. Thanks also to Guy Froyen (VIB, KU Leuven),
Richard Rocca (independent researcher), Cuiping Pan (Stanford University) and
Andreas Keller (Saarland University) for providing us yet unpublished and
published called SNPs of whole genome sequencing projects. Maarten H.D.
Larmuseau is postdoctoral fellow of the FWO-Vlaanderen (Research
Foundation Flanders). This study was funded by the KU Leuven BOF Centre of
Excellence Financing on 'Eco and socio evolutionary dynamics' (Project
number PF/2010/07) and on 'Centre for Archaeological Sciences 2 (CAS 2) New methods for research in demography and interregional exchange'.
Authors’ contributions
Research design & supervision: MHDL; Programming: AVG; Writing: MHDL
& AVG; Commenting on manuscript: AVG, RD & MHDL.
Conflict of interest
The authors declare no conflict of interest.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
M. Kayser, Y-chromosomal markers in forensic genetics, in: R. Rapley and W. D., Editors,
Molecular Forensics, John Wiley & Sons Ltd: Chichesters. 2007, pp. 141-161.
J.M. Butler, Chapter 13 Y-Chromosomal DNA Testing, in: J.M. Butler, Editor, Advanced Topics
in Forensic DNA Typing: Methodology, Academic Press: London. 2011, pp. 371-403.
J. Chiaroni, P.A. Underhill, and L.L. Cavalli-Sforza, Y chromosome diversity, human expansion,
drift, and cultural evolution, Proceedings of the National Academy of Sciences of the United
States of America. 106 (2009) 20174-20179.
T.M. Karafet, F.L. Mendez, M.B. Meilerman, P.A. Underhill, S.L. Zegura, and M.F. Hammer,
New binary polymorphisms reshape and increase resolution of the human Y chromosomal
haplogroup tree, Genome Research. 18 (2008) 830-838.
M.H.D. Larmuseau, N. Vanderheyden, M. Jacobs, M. Coomans, L. Larno, and R. Decorte,
Micro-geographic distribution of Y-chromosomal variation in the central-western European
region Brabant, Forensic Science International-Genetics. 5 (2011) 95-99.
F. Cruciani, B. Trombetta, C. Antonelli, R. Pascone, G. Valesini, V. Scalzi, G. Vona, B. Melegh,
B. Zagradisnik, G. Assum, et al., Strong intra- and inter-continental differentiation revealed
by Y chromosome SNPs M269, U106 and U152, Forensic Science International-Genetics. 5
(2011) E49-E52.
M.H.D. Larmuseau, J. Vanoverbeke, G. Gielis, N. Vanderheyden, H.F.M. Larmuseau, and R.
Decorte, In the name of the migrant father - Analysis of surname origin identifies historic
admixture events undetectable from genealogical records., Heredity. 109 (2012) 90-95.
S. Willuweit and L. Roewer, Y chromosome haplotype reference database (YHRD): Update,
Forensic Science International-Genetics. 1 (2007) 83-87.
R. Scozzari, A. Massaia, E. D'Atanasio, N.M. Myres, U.A. Perego, B. Trombetta, and F.
Cruciani, Molecular dissection of the basal clades in the human Y chromosome phylogenetic
tree, Plos One. 7 (2012) e49170.
F. Cruciani, B. Trombetta, A. Massaia, G. Destro-Bisol, D. Sellitto, and R. Scozzari, A revised
root for the human Y chromosomal phylogenetic tree: The origin of patrilineal diversity in
Africa, American Journal of Human Genetics. 88 (2011) 814-818.
S. Fornarino, M. Pala, V. Battaglia, R. Maranta, A. Achilli, G. Modiano, A. Torroni, O. Semino,
and S.A. Santachiara-Benerecetti, Mitochondrial and Y-chromosome diversity of the Tharus
(Nepal): a reservoir of genetic variation, Bmc Evolutionary Biology. 9 (2009) 154.
F.L. Mendez, T.M. Karafet, T. Krahn, H. Ostrer, H. Soodyall, and M.F. Hammer, Increased
resolution of Y chromosome haplogroup T defines relationships among populations of the
Near East, Europe, and Africa, Human Biology. 83 (2011) 39-53.
L.M. Sims, D. Garvey, and J. Ballantyne, Improved resolution haplogroup G phylogeny in the
Y-chromosome, revealed by a set of newly characterized SNPs, Plos One. 4 (2009) e5792.
B. Trombetta, F. Cruciani, D. Sellitto, and R. Scozzari, A new topology of the human Y
chromosome haplogroup E1b1 (E-P2) revealed through the use of newly characterized binary
polymorphisms, PLoS One. 6 (2011) e16073.
M.S. Jota, D.R. Lacerda, J.R. Sandoval, P.P.R. Vieira, S.S. Santos-Lopes, R. Bisso-Machado, V.R.
Paixao-Cortes, S. Revollo, C. Paz-Y-Mino, R. Fujita, et al., A new subhaplogroup of native
American Y-chromosomes from the Andes, American Journal of Physical Anthropology. 146
(2011) 553-559.
H. Pamjav, T. Feher, E. Nemeth, and Z. Padar, Brief communication: New Y-chromosome
binary markers improve phylogenetic resolution within haplogroup R1a1, American Journal
of Physical Anthropology. 149 (2012) 611-615.
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
N.M. Myres, S. Rootsi, A.A. Lin, M. Jarve, R.J. King, I. Kutuev, V.M. Cabrera, E.K.
Khusnutdinova, A. Pshenichnov, B. Yunusbayev, et al., A major Y-chromosome haplogroup
R1b Holocene era founder effect in Central and Western Europe, European Journal of Human
Genetics. 19 (2011) 95-101.
S. Yan, C.C. Wang, H. Li, S.L. Li, L. Jin, and G. Consortium, An updated tree of Y-chromosome
Haplogroup O and revised phylogenetic positions of mutations P164 and PK4, European
Journal of Human Genetics. 19 (2011) 1013-1015.
T. 1000 Genomes Project Consortium, An integrated map of genetic variation from 1,092
human genomes, Nature. 491 (2012) 56-65.
D.L. Altshuler, R.M. Durbin, G.R. Abecasis, D.R. Bentley, A. Chakravarti, A.G. Clark, F.S.
Collins, F.M. De la Vega, P. Donnelly, M. Egholm, et al., A map of human genome variation
from population-scale sequencing, Nature. 467 (2010) 1061-1073.
A.T. Duggan and M. Stoneking, A highly unstable recent mutation in human mtDNA,
American Journal of Human Genetics. 9 (2013) 279-284.
W. Wei, Q. Ayub, Y. Chen, S. McCarthy, Y. Hou, I. Carbone, Y. Xue, and C. Tyler-Smith, A
calibrated human Y-chromosomal phylogeny based on resequencing, Genome Research. 23
(2013) 388-395.
A. Van Geystelen, R. Decorte, and M.H.D. Larmuseau, AMY-tree: an algorithm to use whole
genome SNP calling for Y chromosomal phylogenetic applications, BMC Genomics. 14 (2013)
101.
R. Drmanac, A.B. Sparks, M.J. Callow, A.L. Halpern, N.L. Burns, B.G. Kermani, P. Carnevali, I.
Nazarenko, G.B. Nilsen, G. Yeung, et al., Human genome sequencing using unchained base
reads on self-assembling DNA nanoarrays, Science. 327 (2010) 78-81.
L.-P. Wong, R.T.-H. Ong, W.-T. Poh, X. Liu, P. Chen, R.Q. Li, K.K.-Y. Lam, N.E. Pillai, K.-S. Sim, H.
Xu, et al., Deep whole-genome sequencing of 100 Southeast Asian Malays, American Journal
of Human Genetics. 92 (2013) 1-15.
S.C. Schuster, W. Miller, A. Ratan, L.P. Tomsho, B. Giardine, L.R. Kasson, R.S. Harris, D.C.
Petersen, F.Q. Zhao, J. Qi, et al., Complete Khoisan and Bantu genomes from southern Africa,
Nature. 463 (2010) 943-947.
P. Tong, J.G.D. Prendergast, A.J. Lohan, S.M. Farrington, S. Cronin, N. Friel, D.G. Bradley, O.
Hardiman, A. Evans, J.F. Wilson, et al., Sequencing and analysis of an Irish human genome,
Genome Biology. 11 (2010) R91.
A. Keller, A. Graefen, M. Ball, M. Matzas, V. Boisguerin, F. Maixner, P. Leidinger, C. Backes, R.
Khairat, M. Forster, et al., New insights into the Tyrolean Iceman's origin and phenotype as
inferred by whole-genome sequencing, Nature communications. 3 (2012) 698.
R. Chen, G.I. Mias, J. Li-Pook-Than, L.H. Jiang, H.Y.K. Lam, R. Chen, E. Miriami, K.J. Karczewski,
M. Hariharan, F.E. Dewey, et al., Personal omics profiling reveals dynamic molecular and
medical phenotypes, Cell. 148 (2012) 1293-1307.
J.M. Rothberg, W. Hinz, T.M. Rearick, J. Schultz, W. Mileski, M. Davey, J.H. Leamon, K.
Johnson, M.J. Milgrew, M. Edwards, et al., An integrated semiconductor device enabling nonoptical genome sequencing, Nature. 475 (2011) 348-352.
D. Pushkarev, N.F. Neff, and S.R. Quake, Single-molecule sequencing of an individual human
genome, Nature Biotechnology. 27 (2009) 847-850.
M. Rasmussen, Y.R. Li, S. Lindgreen, J.S. Pedersen, A. Albrechtsen, I. Moltke, M. Metspalu, E.
Metspalu, T. Kivisild, R. Gupta, et al., Ancient human genome sequence of an extinct PalaeoEskimo, Nature. 463 (2010) 757-762.
S.M. Ahn, T.H. Kim, S. Lee, D. Kim, H. Ghang, D.S. Kim, B.C. Kim, S.Y. Kim, W.Y. Kim, C. Kim, et
al., The first Korean genome sequence and analysis: Full genome sequencing for a socioethnic group, Genome Research. 19 (2009) 1622-1629.
D.A. Wheeler, M. Srinivasan, M. Egholm, Y. Shen, L. Chen, A. McGuire, W. He, Y.J. Chen, V.
Makhijani, G.T. Roth, et al., The complete genome of an individual by massively parallel DNA
sequencing, Nature. 452 (2008) 872-U5.
J. Wang, W. Wang, R.Q. Li, Y.R. Li, G. Tian, L. Goodman, W. Fan, J.Q. Zhang, J. Li, J.B. Zhang, et
al., The diploid genome sequence of an Asian individual, Nature. 456 (2008) 60-U1.
B.A. Peters, B.G. Kermani, A.B. Sparks, O. Alferov, P. Hong, A. Alexeev, Y. Jiang, F. Dahl, Y.T.
Tang, J. Haas, et al., Accurate whole-genome sequencing and haplotyping from 10 to 20
human cells, Nature. 487 (2012) 190-195.
H. Skaletsky, T. Kuroda-Kawaguchi, P.J. Minx, H.S. Cordum, L. Hillier, L.G. Brown, S. Repping,
T. Pyntikova, J. Ali, T. Bieri, et al., The male-specific region of the human Y chromosome is a
mosaic of discrete sequence classes, Nature. 423 (2003) 825-U2.
S.M. Adams, T.E. King, E. Bosch, and M.A. Jobling, The case of the unreliable SNP: recurrent
back-mutation of Y-chromosomal markers P25 through gene conversion, Forensic Science
International. 159 (2006) 14-20.
B. Trombetta, F. Cruciani, P.A. Underhill, D. Sellitto, and R. Scozzari, Footprints of X-to-Y gene
conversion in recent human evolution, Molecular Biology and Evolution. 27 (2010) 714-725.
S. Rootsi, N.M. Myres, A.A. Lin, M. Järve, R.J. King, I. Kutuev, V.M. Cabrera, E.K.
Khusnutdinova, K. Varendi, H. Sahakyan, et al., Distinguishing the co-ancestries of
[41]
[42]
[43]
[44]
[45]
[46]
[47]
haplogroup G Y-chromosomes in the populations of Europe and the Caucasus, European
Journal of Human Genetics. 20 (2012) 1275-1282.
S. Caratti, S. Gino, C. Torre, and C. Robino, Subtyping of Y-chromosomal haplogroup E-M78
(E1b1b1a) by SNP assay and its forensic application, International Journal of Legal Medicine.
123 (2009) 357-360.
C. Bouakaze, C. Keyser, S. Amory, E. Crubézy, and B. Ludes, First successful assay of Y-SNP
typing by SNaPshot minisequencing on ancient DNA, International Journal of Legal Medicine.
121 (2007) 493-499.
Y.L. Xue, Q.J. Wang, Q. Long, B.L. Ng, H. Swerdlow, J. Burton, C. Skuce, R. Taylor, Z. Abdellah,
Y.L. Zhao, et al., Human Y chromosome base-substitution mutation rate measured by direct
sequencing in a deep-rooting pedigree, Current Biology. 19 (2009) 1453-1457.
Y. Kuroki, A. Toyoda, H. Noguchi, T.D. Taylor, T. Itoh, D.S. Kim, D.W. Kim, S.H. Choi, I.C. Kim,
and H.H. Choi, Comparative analysis of chimpanzee and human Y chromosomes unveils
complex evolutionary pathway, Nature Genetics. 38 (2006) 158-167.
T.E. King and M.A. Jobling, What's in a name? Y chromosomes, surnames and the genetic
genealogy revolution, Trends in Genetics. 25 (2009) 351-360.
M.H.D. Larmuseau, A. Van Geystelen, M. van Oven, and R. Decorte, Genetic genealogy comes
of age - Perspectives on the use of deep-rooted pedigrees in human population genetics.,
American Journal of Physical Anthropology. (In press).
M. van Oven and M. Kayser, Updated comprehensive phylogenetic tree of global human
mitochondrial DNA variation, Human Mutation. 30 (2008) E386-E394.
Figures
Select Y-SNPs
based on determined haplogroup
Expected state
ancestral/mutant
State in
reference genome
Expected situation in sample
Actual situation in sample
called/not called
called/not called
TP, FP, TN, FN
Measures
Figure 1. Workflow of the quality assessment algorithm of the AMY-tree,
version 1.1. First, certain Y-SNPs of the Y-chromosomal phylogenetic tree are
selected based on the determined haplogroup of the first run in order to avoid
too much false positive Y-SNPs. Next, the expected state (ancestral or mutant)
of the selected SNPs is determined based on the haplogroup and the
phylogenetic tree. These expectations of state are converted to expectations of
called or not called based on the SNP state in the reference genome. These
expectations are then compared to the actually called SNPs of the sample such
that the number of true positives (TP), false positives (FP), true negatives (TN)
and false negatives (FN) will be determined. Finally, several different measures
will be calculated.
M198, M417, M512, M514, M515, Page7
R-M198*
R-M56
M157.1
R-M157.1
M87, M204
R-M87
P98
R-P98
PK5
R-PK5
M434
R-M434
M458
R-M458*
M334
R-M334
Page68
R-Page68
Z280
R-Z280
Z93
R-Z93
M56
Figure 2. Overview of the position of the newly added sub-haplogroups RZ280 and R-Z93 (given in bold) within R-M198.
Number of samples
frequency
160
80
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Matthew's Correlation
Matthew’s
correlationCoefficient
coefficient
Figure 3. Distribution of the Matthews correlation coefficient of the 730
samples for which an unambiguous haplogroup could be determined.
0.9
1.0
Number of available WGS genomes
165
All genomes
110
55
Good quality
genomes
0
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
Haplogroups
Figure 4. Frequency of whole genome sequencing (WGS) genomes per
haplogroup in the dataset of all samples (black) and in the dataset of samples
with an excellent quality, i.e. MCC ≥ 0.95 (grey).
S
T
70000
A
35000
Number of Y-SNPs
Whole Y-chromosome
70000
All genomes
0
1
2
35000
3
4
5
6
Excellent quality genomes
7
8
9
10
more
9
10
more
Number of occurrences of Y-SNP
0
1
B
2
3
4
5
6
7
8
Unique regions Y-chromosome
24000
Number of occurrences of Y-SNP
12000
24000
0
12000
1
2
3
4
5
6
7
8
9
10
more
Number of occurrences of Y-SNP
Figure 5. Number
of new Y-SNPs per number of occurrence in the full WGS
0
dataset: (A) SNPs 1in the2 whole
Y-chromosome
and7(B) SNPs
in only
the
3
4
5
6
8
9
10 more
identified unique regions of the Y-chromosome. The grey bars indicate the
SNPs in all genomes and the black bars
indicate
the SNPs
Number
of occurrences
of from
Y-SNPsamples with an
excellent quality, i.e. MCC ≥ 0.95.
2500
2000
A
1500 of Y-SNPs
Number
Whole Y-chromosome
1000
2500
500
2000
All SNPs
0
1500
Family-unique SNPs
1000
500
0
8-member family
B
Father-son pair
Unique regions Y-chromosome
120
100
80
60
120
40
100
20
80
0
60
1
2
3
4
5
6
7
8
1
2
40
20
Number of occurrences of Y-SNP
0
Figure 6. Number of
1 new2Y-SNPs
3 per4 occurrence
5
6in the7 eight8paternally related
1
2
samples and the father-son pair of Complete Genomics: (A) SNPs in the whole
Y-chromosome and (B) SNPs in only
the ofidentified
unique
regions of the YNumber
occurrences
of Y-SNP
chromosome. The grey bars indicate the SNPs found in the family. The black
bars indicate the SNPs which are ‘family-unique’: they do not occur in any of
the other samples with an excellent quality, i.e. MCC ≥ 0.95.