Statistical estimation of cell-cycle progression and lineage commitment in Plasmodium falciparum reveals a homogeneous pattern of transcription in ex vivo culture

Abstract

We have cultured Plasmodium falciparum directly from the blood of infected individuals to examine patterns of mature-stage gene expression in patient isolates. Analysis of the transcriptome of P. falciparum is complicated by the highly periodic nature of gene expression because small variations in the stage of parasite development between samples can lead to an apparent difference in gene expression values. To address this issue, we have developed statistical likelihood-based methods to estimate cell cycle progression and commitment to asexual or sexual development lineages in our samples based on microscopy and gene expression patterns. In cases subsequently matched for temporal development, we find that transcriptional patterns in ex vivo culture display little variation across patients with diverse clinical profiles and closely resemble transcriptional profiles that occur in vitro. These statistical methods, available to the research community, assist in the design and interpretation of P. falciparum expression profiling experiments where it is difficult to separate true differential expression from cell-cycle dependent expression. We reanalyze an existing dataset of in vivo patient expression profiles and conclude that previously observed discrete variation is consistent with the commitment of a varying proportion of the parasite population to the sexual development lineage.

Fig. 1.

Estimating parasite age. (A) Estimates of parasite age from gene expression based on a reference set correlate with measurements of parasite area. (B) The maximum likelihood estimates and 95% confidence intervals for the samples in this study are shown. These estimates span the interval from 20 HPI to 44 HPI. Samples are ordered in the y dimension according to maximum likelihood estimate. (C) Each curve corresponds to a single sample and represents a probabilistic interpretation of how likely the sample was to have come from a reference value at a given point in the lifecycle. The curves are constructed by obtaining the product of the likelihood densities for each individual gene that matches between the test and reference set. The maxima and shape of these curves are used to produce the maximum likelihood estimate and confidence interval, respectively, for each sample. (D) A normalized heatmap of the log-likelihood curves ordered by maximum. Each sample is a row in the heatmap, and the x dimension corresponds to hours after invasion. The height of the log-likelihood curve at each point is displayed as color.

Fig. 2.

Maximum likelihood estimates (A) and log-likelihood curves (B) for samples from Daily et al. (2). (A) The samples all have maxima within the ring portion of the parasite lifecycle. The clusters do not appear to be associated with different sample age. (B) The log-likelihood curves show a continuous progression away from the ring signal, which corresponds to an increase in log-likelihood at ≈30 h, suggestive of the contribution of gametocyte gene expression patterns. (C) The gene expression profiles of the samples grown ex vivo, along with in vitro controls from Cortes et al. (17), were projected into the space defined by the first 2 principal components of the gametocyte development time course. The samples from this study project in a distribution centered around in vitro late-stage schizonts (triangles), indicating that there is little sexual-specific transcription occurring in these samples. (D) The samples from the study by Daily et al. (2) follow a continuous progression into the gametocyte development space, and are not distributed around time-matched ring stage samples or trophozoite samples from the 3D7 strain (6) measured on the same (scrmalaria) GeneChip array. Clusters identified by Daily et al. (2) are associated with the depth of projection in the sexual development space. (E) Maximum likelihood estimates of sample age (x axis) and lineage commitment (y axis). The ex vivo samples in this study are distributed along the late stages of the asexual development cycle with small values of α, the fraction of the culture committed to sexual development. In vitro asexual stages (6, 17) and early days 1–3 of mixed gametocyte/asexual cultures (20) are plotted as controls. The in vivo samples from Daily et al. (2) are narrowly distributed in the early ring stage but widely and almost uniformly across values of α. The observed clusters from Daily et al. (2) are associated with different values of this parameter.