DNA barcoding of fungi: a feasibility analysis

Donal Hickey,

Concordia University,

Montreal, Canada

2 of 20

General questions:Is DNA barcoding a taxonomic tool?

- a phylogenetic tool?

- a tool for simply assigning unidentified specimens to known species?

- all of the above?

- none of the above?

Specific question:

Will DNA barcoding work for fungi?

The big question:Should we be using mitochondrial sequences DNA barcoding?

3 of 20

First, let’s see how barcoding works in a case where we know the answer.

Then we can move on, with more confidence, to cases such as fungi – where we definitely don’t know the answer.

4 of 20

The CBOL brochure:

What is wrong with this picture?

5 of 20

Benchmarking DNA barcodes: an assessment using available

primate sequences

6 of 20

By using high bootstrap values, most of the branching pattern collapses, but the species resolution remains.

7 of 20

The two closely related species of chimpanzee can be resolved by reducing the bootstrap cut-off to 95%

8 of 20

Or, we can retain the 100% bootstrap and increase the sequence length – it’s a simple trade-off .

In this case the “barcode” was extended to 1,500 bp

9 of 20

10 of 20

Let’s begin with long sequences (5 concatenated genes)

11 of 20

Then, let’s use a single genes (CO1)

12 of 20

How about Cytochrome b?

13 of 20

CO1 Barcode (600 bp)

14 of 20

Multiple strains within a species

15 of 20

The relationship between barcode length and diagnostic value

(Lepidopteran dataset)

16 of 20

Barcodes for genome composition :GC content of animal mitochondria

17 of 20

Are mitochondrial barcodes a bad idea?

18 of 20

Count the green dots

19 of 20

Numbers of mitochondrial genomes in the mammalian female germ line fluctuate between 1,000,000 per cell and 100 per cell.

But the “bottleneck” in a single female is approximately 5,000 mit genomes.

(Tim Wai, McGill)

20 of 20

The bad news:

Mitochondrial sequence variation does not reflect the population size and/or the breeding pattern of the population.

“Panmictic population” has no meaning when applied to mitochondrial data.

21 of 20

The good news:

Since mitochondrial genes do not reflect the effects of sexual outbreeding, mitochondrial barcodes should work equally well in populations with different breeding structures.

In other words, we shouldn’t worry too much about the fact that fungi, unlike birds, do not come in breeding pairs.

22 of 20

The cautionary note.

Since different lineages have:

(i) different cell sizes, correlated with different numbers of mitochondria per cell; and

(ii) different numbers of cell generations per organismal generation,

we should expect to see large variations in the relative rates of mitochondrial and nuclear sequence evolution, even if the base mutation rates were the same.

23 of 20

A modest proposal for a barcoder’s credo

1. Leave phylogenetics to the phylogeneticists (i.e., use their trees)

2. Leave taxonomy to the taxonomists (i.e. confine ourselves to their Latin binomial names).

3. Assign barcode sequences to known species names and hang them on independently derived trees (otherwise, we are crossing a line into DNA taxonomy and molecular phylogenetics).

24 of 20

Visualization: the color scheme.

Perfect match to verified voucher specimen: GREEN

Perfect match to known species, but no verified voucher with exactly that barcode sequence: BLACK

<0.25% mismatch to known sequence/specimen: YELLOW

<0.50% mismatch to known sequence/specimen: ORANGE

<0.75% mismatch to known sequence/specimen: RED

<1.00% mismatch to known sequence/specimen: RED

>1.00% mismatch to known sequence: PROBLEM!

(seek help from a qualified professional)

25 of 20

A field test

26 of 20

*

The field test

Duhamel

27 of 20

Acknowledgements

Xiang Jia Min

Mehrdad Hajibabaei

Greg Singer

The Canadian Taxpayer