Fig 1.
Sketch of stochastic processes involved in data generation process.
(A) The four processes indicated by host 1 infecting host 2, together leading to a difference between the sampled sequences of hosts 1 and 2: (a) transmission, (b) sampling, (c) coalescence, and (d) mutation. As described in Methods, the node ID numbers are indicated in circles as if this was a complete outbreak. (B) Examples of differences in sequences for host a infecting both hosts 2 and 3. Host 1 is infected by the original sequence xyz, and the lightning indicates when mutations take place.
Table 1.
Performance of the method: analysis of 25 newly simulated datasets of 50 cases, with shape parameters aS = aG = 10.
Table 2.
Tree inference with incomplete data, with 25 newly simulated datasets of 50 cases, of which 40 observed, simulated with shape parameters aS = aG = 10.
Table 3.
Performance on 25 published simulated datasets in populations of size 50 [12].
Fig 2.
Consensus Edmonds’ transmission trees for four of the five analysed datasets.
Crosses indicate sampling days, coloured links indicate most likely infectors, with colours indicating the posterior support for that infector. (A) Mtb data [7]; (B) MRSA data [25]; (C) FMD2001 data [26]; (D) FMD2007 data [27].
Fig 3.
Consensus MPC transmission and phylogenetic trees for four of the five analysed datasets.
Each tree is one posterior sample matching the MPC tree topology. Colours are used to indicate the hosts in the transmission tree: connected branches with identical colour are in the same host, and a change of colour along a branch is a transmission event. (A) Mtb data [7]; (B) MRSA data [25]; (C) FMD2001 data [9, 26]; (D) FMD2007 data [9, 27].
Table 4.
Summary statistics for four published datasets.
Fig 4.
Consensus Edmonds’ transmission tree for the H7N7 dataset [12, 28, 30].
Infected premises are (not uniquely) coded by location (as in [12]), median posterior infection day, and sampling day. Coloured arrows indicate most likely infectors, with colours indicating the posterior support for that infector. The asterisk (*) indicates the exceptionally long generation time at the start of a small Limburg cluster.
Table 5.
Summary statistics for H7N7 dataset.
Fig 5.
Graphics depicting proposal steps A-F for new transmission and phylogenetic trees. In panels A, B, D, and E, the initial situation is at the top, and the proposal below. In panels C and F, the initial situation is in the middle, and two alternative proposal above and below. Every panel shows an outbreak with four hosts, with red arrows indicating transmission: the purple host is the focal host, with the purple arrow indicating the proposal for the new infection time Ii’; filled hosts have a new phylogenetic mini-tree proposed; greyed-out hosts do not play a role in the proposal. (A) the focal host is the index case, and Ii’ is before the first transmission event; (B) the focal host is the index case, and Ii’ is after the first, but before the second secondary case; (C) the focal host is the index case and Ii’ is after his second secondary case; (D) the focal host is not the index case and Ii’ is before infection of the index case; (E) the focal host is not the index case and Ii’ is before his first secondary case; (F) the focal host is not the index case and Ii’ is after his first secondary case.
Fig 6.
Graphics depicting proposal steps H-J for new transmission trees, keeping the phylogenetic tree unchanged. In all panels, the initial situation is at the top, and the proposal below. Every panel shows part of an outbreak, with red arrows indicating transmission to depicted or undepicted hosts. Only in panel B host I must be the index case. The purple host is the focal host, with the dark purple arrow indicating the proposal for the new infection time Ii’; the light purple arrow in panels B and C indicate the proposal for the new infection time Ij’ of the focal host’s infector. The grey parts of the phylogenetic tree are moved between the hosts. (H) the focal host is not the index case, and Ii’ is after MRCAI,II of the focal host and his infector; (I) the focal host is not the index case, and Ii’ is before MRCAI,II of the focal host and his infector (the index case), and Ij’ is after MRCAI,II; (J) the focal host is not the index case, and Ii’ is before MRCAII,III of the focal host and his infector, but after the MRCAI,III of the focal host and his infector’s infector; also, Ij’ is after MRCAII,III.